search

HolyWu / vs-femasr

FeMaSR function for VapourSynth
Creative Commons Zero v1.0 Universal
13 stars 0 forks source link

Tensor issue,... #6

Closed Selur closed 5 months ago

Selur commented 1 year ago

Using:

# Imports
import vapoursynth as vs
import os
import ctypes
# Loading Support Files
Dllref = ctypes.windll.LoadLibrary("i:/Hybrid/64bit/vsfilters/Support/libfftw3f-3.dll")
import sys
# getting Vapoursynth core
core = vs.core
# Import scripts folder
scriptPath = 'i:/Hybrid/64bit/vsscripts'
sys.path.insert(0, os.path.abspath(scriptPath))
import site
# Adding torch dependencies to PATH
path = site.getsitepackages()[0]+'/torch_dependencies/bin/'
ctypes.windll.kernel32.SetDllDirectoryW(path)
path = path.replace('\\', '/')
os.environ["PATH"] = path + os.pathsep + os.environ["PATH"]
os.environ["CUDA_MODULE_LOADING"] = "LAZY"
# Loading Plugins
core.std.LoadPlugin(path="i:/Hybrid/64bit/vsfilters/GrainFilter/RemoveGrain/RemoveGrainVS.dll")
core.std.LoadPlugin(path="i:/Hybrid/64bit/vsfilters/GrainFilter/AddGrain/AddGrain.dll")
core.std.LoadPlugin(path="i:/Hybrid/64bit/vsfilters/DenoiseFilter/DFTTest/DFTTest.dll")
core.std.LoadPlugin(path="i:/Hybrid/64bit/vsfilters/DenoiseFilter/FFT3DFilter/fft3dfilter.dll")
core.std.LoadPlugin(path="i:/Hybrid/64bit/vsfilters/Support/EEDI3m_opencl.dll")
core.std.LoadPlugin(path="i:/Hybrid/64bit/vsfilters/ResizeFilter/nnedi3/NNEDI3CL.dll")
core.std.LoadPlugin(path="i:/Hybrid/64bit/vsfilters/Support/libmvtools.dll")
core.std.LoadPlugin(path="i:/Hybrid/64bit/vsfilters/Support/scenechange.dll")
core.std.LoadPlugin(path="i:/Hybrid/64bit/vsfilters/Support/fmtconv.dll")
core.std.LoadPlugin(path="i:/Hybrid/64bit/vsfilters/MiscFilter/MiscFilters/MiscFilters.dll")
core.std.LoadPlugin(path="i:/Hybrid/64bit/vsfilters/DeinterlaceFilter/Bwdif/Bwdif.dll")
core.std.LoadPlugin(path="i:/Hybrid/64bit/vsfilters/SourceFilter/DGDecNV/DGDecodeNV.dll")
# Import scripts
import havsfunc
# source: 'G:\TestClips&Co\1983-i.mkv'
# current color space: YUV420P8, bit depth: 8, resolution: 720x576, fps: 25, color matrix: 470bg, yuv luminance scale: limited, scanorder: top field first
# Loading G:\TestClips&Co\1983-i.mkv using DGSource
clip = core.dgdecodenv.DGSource("J:/tmp/mkv_85eb1ac8734d03d3e24c3850e2e9287a_853323747.dgi",fieldop=0)# 25 fps, scanorder: top field first
# Setting detected color matrix (470bg).
clip = core.std.SetFrameProps(clip, _Matrix=5)
# Setting color transfer info (470bg), when it is not set
clip = clip if not core.text.FrameProps(clip,'_Transfer') else core.std.SetFrameProps(clip, _Transfer=5)
# Setting color primaries info (), when it is not set
clip = clip if not core.text.FrameProps(clip,'_Primaries') else core.std.SetFrameProps(clip, _Primaries=5)
# Setting color range to TV (limited) range.
clip = core.std.SetFrameProp(clip=clip, prop="_ColorRange", intval=1)
# making sure frame rate is set to 25
clip = core.std.AssumeFPS(clip=clip, fpsnum=25, fpsden=1)
clip = core.std.SetFrameProp(clip=clip, prop="_FieldBased", intval=2) # tff
# Deinterlacing using QTGMC
clip = havsfunc.QTGMC(Input=clip, Preset="Fast", TFF=True, opencl=True) # new fps: 50
# Making sure content is preceived as frame based
clip = core.std.SetFrameProp(clip=clip, prop="_FieldBased", intval=0) # progressive
from vsfemasr import femasr as FeMaSR
# adjusting color space from YUV420P8 to RGBH for VsFeMaSR
clip = core.resize.Bicubic(clip=clip, format=vs.RGBH, matrix_in_s="470bg", range_s="limited")
# resizing using FeMaSR
clip = FeMaSR(clip=clip, device_index=0, trt=True, trt_cache_path=r"J:\tmp") # 1440x1152
# resizing 1440x1152 to 1280x960
# adjusting resizing
clip = core.resize.Bicubic(clip=clip, format=vs.RGBS, range_s="limited")
clip = core.fmtc.resample(clip=clip, w=1280, h=960, kernel="lanczos", interlaced=False, interlacedd=False)
# adjusting output color from: RGBS to YUV420P10 for x265Model
clip = core.resize.Bicubic(clip=clip, format=vs.YUV420P10, matrix_s="470bg", range_s="limited", dither_type="error_diffusion")
# set output frame rate to 50fps (progressive)
clip = core.std.AssumeFPS(clip=clip, fpsnum=50, fpsden=1)
# Output
clip.set_output()

I get:

<html><head><meta charset="utf-8" /><style type="text/css">
p, li { white-space: pre-wrap; }
hr { height: 1px; border-width: 0; }
li.unchecked::marker { content: "\2610"; }
li.checked::marker { content: "\2612"; }
</style></head><body>

167 entries not shown. Save the log to read.
--
2023-03-26 17:12:10.862 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_568 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_569 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_569 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_576 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_576 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_579 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_579 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_581 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_581 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_582 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_582 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_589 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_589 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_592 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_592 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_594 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_594 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_595 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_595 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_602 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_602 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_605 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_605 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_607 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_607 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_608 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_608 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_615 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_615 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_618 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_618 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_620 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_620 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_621 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_621 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_628 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_628 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_631 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_631 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_633 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_633 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_634 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_634 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_635 : Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_635 Now lowering submodule _run_on_acc_0 Now lowering submodule _run_on_acc_0 split_name=_run_on_acc_0, input_specs=[InputTensorSpec(shape=torch.Size([1, 3, 576, 736]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] split_name=_run_on_acc_0, input_specs=[InputTensorSpec(shape=torch.Size([1, 3, 576, 736]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] Timing cache is used! Timing cache is used! TRT INetwork construction elapsed time: 0:00:00.001000 TRT INetwork construction elapsed time: 0:00:00.001000
2023-03-26 17:12:16.295 Build TRT engine elapsed time: 0:00:02.024681 Build TRT engine elapsed time: 0:00:02.024681 Lowering submodule _run_on_acc_0 elapsed time 0:00:03.686569 Lowering submodule _run_on_acc_0 elapsed time 0:00:03.686569 Now lowering submodule _run_on_acc_2 Now lowering submodule _run_on_acc_2 split_name=_run_on_acc_2, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 288, 368]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] split_name=_run_on_acc_2, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 288, 368]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] Timing cache is used! Timing cache is used! TRT INetwork construction elapsed time: 0:00:00.001505 TRT INetwork construction elapsed time: 0:00:00.001505
2023-03-26 17:12:37.284 Build TRT engine elapsed time: 0:00:20.702133 Build TRT engine elapsed time: 0:00:20.702133 Lowering submodule _run_on_acc_2 elapsed time 0:00:20.896665 Lowering submodule _run_on_acc_2 elapsed time 0:00:20.896665 Now lowering submodule _run_on_acc_4 Now lowering submodule _run_on_acc_4 split_name=_run_on_acc_4, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 288, 368]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 288, 368]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] split_name=_run_on_acc_4, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 288, 368]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 288, 368]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] Timing cache is used! Timing cache is used! TRT INetwork construction elapsed time: 0:00:00.001001 TRT INetwork construction elapsed time: 0:00:00.001001
2023-03-26 17:12:40.478 Build TRT engine elapsed time: 0:00:02.913266 Build TRT engine elapsed time: 0:00:02.913266 Lowering submodule _run_on_acc_4 elapsed time 0:00:03.107299 Lowering submodule _run_on_acc_4 elapsed time 0:00:03.107299 Now lowering submodule _run_on_acc_6 Now lowering submodule _run_on_acc_6 split_name=_run_on_acc_6, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 288, 368]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] split_name=_run_on_acc_6, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 288, 368]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] Timing cache is used! Timing cache is used! TRT INetwork construction elapsed time: 0:00:00.001001 TRT INetwork construction elapsed time: 0:00:00.001001
2023-03-26 17:12:43.662 Build TRT engine elapsed time: 0:00:02.897112 Build TRT engine elapsed time: 0:00:02.897112 Lowering submodule _run_on_acc_6 elapsed time 0:00:03.091014 Lowering submodule _run_on_acc_6 elapsed time 0:00:03.091014 Now lowering submodule _run_on_acc_8 Now lowering submodule _run_on_acc_8 split_name=_run_on_acc_8, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 288, 368]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 288, 368]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] split_name=_run_on_acc_8, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 288, 368]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 288, 368]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] Timing cache is used! Timing cache is used! TRT INetwork construction elapsed time: 0:00:00.002001 TRT INetwork construction elapsed time: 0:00:00.002001
2023-03-26 17:12:47.461 Build TRT engine elapsed time: 0:00:03.509264 Build TRT engine elapsed time: 0:00:03.509264 Lowering submodule _run_on_acc_8 elapsed time 0:00:03.711010 Lowering submodule _run_on_acc_8 elapsed time 0:00:03.711010 Now lowering submodule _run_on_acc_10 Now lowering submodule _run_on_acc_10 split_name=_run_on_acc_10, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 144, 184]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] split_name=_run_on_acc_10, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 144, 184]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] Timing cache is used! Timing cache is used! TRT INetwork construction elapsed time: 0:00:00.002001 TRT INetwork construction elapsed time: 0:00:00.002001 Build TRT engine elapsed time: 0:00:01.452047 Build TRT engine elapsed time: 0:00:01.452047 Lowering submodule _run_on_acc_10 elapsed time 0:00:01.644582 Lowering submodule _run_on_acc_10 elapsed time 0:00:01.644582 Now lowering submodule _run_on_acc_12 Now lowering submodule _run_on_acc_12 split_name=_run_on_acc_12, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 144, 184]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 144, 184]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] split_name=_run_on_acc_12, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 144, 184]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 144, 184]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] Timing cache is used! Timing cache is used! TRT INetwork construction elapsed time: 0:00:00.000999 TRT INetwork construction elapsed time: 0:00:00.000999 Build TRT engine elapsed time: 0:00:01.452022 Build TRT engine elapsed time: 0:00:01.452022 Lowering submodule _run_on_acc_12 elapsed time 0:00:01.650969 Lowering submodule _run_on_acc_12 elapsed time 0:00:01.650969 Now lowering submodule _run_on_acc_14 Now lowering submodule _run_on_acc_14 split_name=_run_on_acc_14, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 144, 184]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] split_name=_run_on_acc_14, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 144, 184]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] Timing cache is used! Timing cache is used! TRT INetwork construction elapsed time: 0:00:00.001000 TRT INetwork construction elapsed time: 0:00:00.001000 Build TRT engine elapsed time: 0:00:01.454924 Build TRT engine elapsed time: 0:00:01.454924 Lowering submodule _run_on_acc_14 elapsed time 0:00:01.652692 Lowering submodule _run_on_acc_14 elapsed time 0:00:01.652692 Now lowering submodule _run_on_acc_16 Now lowering submodule _run_on_acc_16 split_name=_run_on_acc_16, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 144, 184]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 144, 184]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] split_name=_run_on_acc_16, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 144, 184]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 144, 184]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] Timing cache is used! Timing cache is used! TRT INetwork construction elapsed time: 0:00:00.001000 TRT INetwork construction elapsed time: 0:00:00.001000 Build TRT engine elapsed time: 0:00:01.471268 Build TRT engine elapsed time: 0:00:01.471268 Lowering submodule _run_on_acc_16 elapsed time 0:00:01.675096 Lowering submodule _run_on_acc_16 elapsed time 0:00:01.675096 Now lowering submodule _run_on_acc_18 Now lowering submodule _run_on_acc_18 split_name=_run_on_acc_18, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 26496]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] split_name=_run_on_acc_18, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 26496]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] Timing cache is used! Timing cache is used!
2023-03-26 17:12:54.710 Unable to find layer norm plugin, fall back to TensorRT implementation. Unable to find layer norm plugin, fall back to TensorRT implementation.
2023-03-26 17:12:54.761 TRT INetwork construction elapsed time: 0:00:00.050510 TRT INetwork construction elapsed time: 0:00:00.050510
2023-03-26 17:13:16.078 Build TRT engine elapsed time: 0:00:21.268365 Build TRT engine elapsed time: 0:00:21.268365 Lowering submodule _run_on_acc_18 elapsed time 0:00:21.520086 Lowering submodule _run_on_acc_18 elapsed time 0:00:21.520086 Now lowering submodule _run_on_acc_20 Now lowering submodule _run_on_acc_20 split_name=_run_on_acc_20, input_specs=[InputTensorSpec(shape=torch.Size([1, 144, 184, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 144, 184, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] split_name=_run_on_acc_20, input_specs=[InputTensorSpec(shape=torch.Size([1, 144, 184, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 144, 184, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] Timing cache is used! Timing cache is used! I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_144 are constant. In this case, please consider constant fold the model first.warnings.warn( I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_144 are constant. In this case, please consider constant fold the model first.warnings.warn( I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_145 are constant. In this case, please consider constant fold the model first.warnings.warn( I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_145 are constant. In this case, please consider constant fold the model first.warnings.warn( I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_146 are constant. In this case, please consider constant fold the model first.warnings.warn( I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_146 are constant. In this case, please consider constant fold the model first.warnings.warn( I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_147 are constant. In this case, please consider constant fold the model first.warnings.warn( I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_147 are constant. In this case, please consider constant fold the model first.warnings.warn( TRT INetwork construction elapsed time: 0:00:00.226765 TRT INetwork construction elapsed time: 0:00:00.226765 Build TRT engine elapsed time: 0:00:00.463744 Build TRT engine elapsed time: 0:00:00.463744 Lowering submodule _run_on_acc_20 elapsed time 0:00:00.918417 Lowering submodule _run_on_acc_20 elapsed time 0:00:00.918417 Now lowering submodule _run_on_acc_22 Now lowering submodule _run_on_acc_22 split_name=_run_on_acc_22, input_specs=[InputTensorSpec(shape=torch.Size([414, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([414, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 414, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] split_name=_run_on_acc_22, input_specs=[InputTensorSpec(shape=torch.Size([414, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([414, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 414, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)] Timing cache is used! Timing cache is used! I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_148 are constant. In this case, please consider constant fold the model first.warnings.warn( I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_148 are constant. In this case, please consider constant fold the model first.warnings.warn( I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_149 are constant. In this case, please consider constant fold the model first.warnings.warn( I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_149 are constant. In this case, please consider constant fold the model first.warnings.warn( TRT INetwork construction elapsed time: 0:00:00.126577 TRT INetwork construction elapsed time: 0:00:00.126577
2023-03-26 17:13:22.295 Failed to evaluate the script:Python exception: Traceback (most recent call last):File "src\cython\vapoursynth.pyx", line 2866, in vapoursynth._vpy_evaluateFile "src\cython\vapoursynth.pyx", line 2867, in vapoursynth._vpy_evaluateFile "J:\tmp\tempPreviewVapoursynthFile17_12_01_348.vpy", line 63, in clip = FeMaSR(clip=clip, device_index=0, trt=True, trt_cache_path=r"J:\tmp") # 1440x1152File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch\autograd\grad_mode.py", line 27, in decorate_contextreturn func(*args, **kwargs)File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\vsfemasr\__init__.py", line 171, in femasrmodule = lowerer(File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\lower.py", line 323, in __call__return do_lower(module, inputs)File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\passes\pass_utils.py", line 117, in pass_with_validationprocessed_module = pass_(module, input, *args, **kwargs)File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\lower.py", line 320, in do_lowerlower_result = pm(module)File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch\fx\passes\pass_manager.py", line 240, in __call__out = _pass(out)File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch\fx\passes\pass_manager.py", line 240, in __call__out = _pass(out)File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\passes\lower_pass_manager_builder.py", line 167, in lower_funclowered_module = self._lower_func(File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\lower.py", line 180, in lower_passinterp_res: TRTInterpreterResult = interpreter(mod, input, module_name)File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\lower.py", line 132, in __call__interp_result: TRTInterpreterResult = interpreter.run(File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\fx2trt.py", line 252, in runassert engineAssertionError

</body></html><!--EndFragment-->

with trt=False it works. Using trt_min_subgraph_size=5 doesn't help.

Any idea what could be causing this? (updated to NIVIDA Studio Drivers 531.41 a few days ago, could this be the cause?)

HolyWu commented 1 year ago

Your error messages are hard to view and have html garbage code. Anyway, it's caused by insufficient VRAM when building the TensorRT engine. You can try a small resolutions like 320x240 and see if it builds fine. If larger resolutions fail to build the engine then just use nvFuser + CUDA Graphs.

Selur commented 1 year ago

Strange, I just tried

# Imports
import vapoursynth as vs
# getting Vapoursynth core
core = vs.core
import site
import os
import ctypes
# Adding torch dependencies to PATH
path = site.getsitepackages()[0]+'/torch_dependencies/bin/'
ctypes.windll.kernel32.SetDllDirectoryW(path)
path = path.replace('\\', '/')
os.environ["PATH"] = path + os.pathsep + os.environ["PATH"]
os.environ["CUDA_MODULE_LOADING"] = "LAZY"
# Loading Plugins
core.std.LoadPlugin(path="i:/Hybrid/64bit/vsfilters/Support/fmtconv.dll")
core.std.LoadPlugin(path="i:/Hybrid/64bit/vsfilters/SourceFilter/LSmashSource/vslsmashsource.dll")
# source: 'G:\TestClips&Co\test.avi'
# current color space: YUV420P8, bit depth: 8, resolution: 640x352, fps: 25, color matrix: 470bg, yuv luminance scale: limited, scanorder: progressive
# Loading G:\TestClips&Co\test.avi using LWLibavSource
clip = core.lsmas.LWLibavSource(source="G:/TestClips&Co/test.avi", format="YUV420P8", stream_index=0, cache=0, prefer_hw=0)
# Setting detected color matrix (470bg).
clip = core.std.SetFrameProps(clip, _Matrix=5)
# Setting color transfer info (470bg), when it is not set
clip = clip if not core.text.FrameProps(clip,'_Transfer') else core.std.SetFrameProps(clip, _Transfer=5)
# Setting color primaries info (), when it is not set
clip = clip if not core.text.FrameProps(clip,'_Primaries') else core.std.SetFrameProps(clip, _Primaries=5)
# Setting color range to TV (limited) range.
clip = core.std.SetFrameProp(clip=clip, prop="_ColorRange", intval=1)
# making sure frame rate is set to 25
clip = core.std.AssumeFPS(clip=clip, fpsnum=25, fpsden=1)
clip = core.std.SetFrameProp(clip=clip, prop="_FieldBased", intval=0) # progressive
from vsfemasr import femasr as FeMaSR
# adjusting color space from YUV420P8 to RGBH for VsFeMaSR
clip = core.resize.Bicubic(clip=clip, format=vs.RGBH, matrix_in_s="470bg", range_s="limited")
# resizing using FeMaSR
clip = FeMaSR(clip=clip, device_index=0, trt=True, trt_cache_path=r"J:\tmp") # 1280x704
# resizing 1280x704 to 640x352
# adjusting resizing
clip = core.resize.Bicubic(clip=clip, format=vs.RGBS, range_s="limited")
clip = core.fmtc.resample(clip=clip, w=640, h=352, kernel="lanczos", interlaced=False, interlacedd=False)
# adjusting output color from: RGBS to YUV420P10 for x265Model
clip = core.resize.Bicubic(clip=clip, format=vs.YUV420P10, matrix_s="470bg", range_s="limited", dither_type="error_diffusion")
# set output frame rate to 25fps (progressive)
clip = core.std.AssumeFPS(clip=clip, fpsnum=25, fpsden=1)
# Output
clip.set_output()

and monitored the VRAM usage and it only got up to 8.4 of 16GB (using a Geforce RTX 4080) for a short time and manly was around 7GB, I still got:

2023-03-26 19:54:49.029

Skip rewriting leaf module

Skip rewriting leaf module

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\tracer\acc_tracer\acc_tracer.py:584: UserWarning: acc_tracer does not support currently support models for training. Calling eval on model before tracing.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\tracer\acc_tracer\acc_tracer.py:584: UserWarning: acc_tracer does not support currently support models for training. Calling eval on model before tracing.
warnings.warn(

Skip rewriting leaf module

Skip rewriting leaf module

== Log pass before/after graph to C:\Users\Selur\AppData\Local\Temp\tmpffq7z0qh, before/after are the same = False

== Log pass before/after graph to C:\Users\Selur\AppData\Local\Temp\tmpffq7z0qh, before/after are the same = False

== Log pass before/after graph to C:\Users\Selur\AppData\Local\Temp\tmpcq83atls, before/after are the same = True

== Log pass before/after graph to C:\Users\Selur\AppData\Local\Temp\tmpcq83atls, before/after are the same = True

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_319

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_319

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_326

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_326

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_329

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_329

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_331

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_331

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_332

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_332

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_339

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_339

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_342

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_342

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_344

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_344

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_345

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_345

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_352

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_352

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_355

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_355

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_357

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_357

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_358

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_358

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_365

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_365

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_368

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_368

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_370

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_370

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_371

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_371

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_378

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_378

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_381

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_381

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_383

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_383

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_384

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_384

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_391

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_391

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_394

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_394

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_396

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_396

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_397

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_397

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_398

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_398

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_405

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_405

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_408

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_408

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_410

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_410

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_411

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_411

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_418

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_418

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_421

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_421

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_423

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_423

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_424

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_424

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_431

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_431

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_434

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_434

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_436

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_436

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_437

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_437

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_444

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_444

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_447

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_447

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_449

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_449

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_450

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_450

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_457

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_457

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_460

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_460

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_462

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_462

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_463

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_463

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_470

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_470

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_473

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_473

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_475

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_475

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_476

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_476

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_477

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_477

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_484

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_484

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_487

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_487

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_489

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_489

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_490

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_490

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_497

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_497

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_500

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_500

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_502

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_502

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_503

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_503

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_510

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_510

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_513

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_513

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_515

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_515

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_516

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_516

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_523

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_523

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_526

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_526

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_528

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_528

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_529

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_529

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_536

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_536

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_539

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_539

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_541

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_541

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_542

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_542

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_549

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_549

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_552

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_552

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_554

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_554

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_555

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_555

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_556

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_556

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_563

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_563

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_566

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_566

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_568

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_568

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_569

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_569

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_576

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_576

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_579

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_579

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_581

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_581

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_582

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_582

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_589

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_589

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_592

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_592

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_594

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_594

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_595

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_595

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_602

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_602

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_605

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_605

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_607

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_607

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_608

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_608

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_615

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_615

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_618

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_618

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_620

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_620

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_621

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_621

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_628

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_628

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_631

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_631

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_633

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_633

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_634

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_634

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_635

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_635

Now lowering submodule _run_on_acc_0

Now lowering submodule _run_on_acc_0

split_name=_run_on_acc_0, input_specs=[InputTensorSpec(shape=torch.Size([1, 3, 352, 640]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_0, input_specs=[InputTensorSpec(shape=torch.Size([1, 3, 352, 640]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.001383

TRT INetwork construction elapsed time: 0:00:00.001383

2023-03-26 19:54:58.127

Build TRT engine elapsed time: 0:00:02.936065

Build TRT engine elapsed time: 0:00:02.936065

Lowering submodule _run_on_acc_0 elapsed time 0:00:04.452679

Lowering submodule _run_on_acc_0 elapsed time 0:00:04.452679

Now lowering submodule _run_on_acc_2

Now lowering submodule _run_on_acc_2

split_name=_run_on_acc_2, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_2, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.002001

TRT INetwork construction elapsed time: 0:00:00.002001

2023-03-26 19:55:19.541

Build TRT engine elapsed time: 0:00:21.371700

Build TRT engine elapsed time: 0:00:21.371700

Lowering submodule _run_on_acc_2 elapsed time 0:00:21.400930

Lowering submodule _run_on_acc_2 elapsed time 0:00:21.400930

Now lowering submodule _run_on_acc_4

Now lowering submodule _run_on_acc_4

split_name=_run_on_acc_4, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_4, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.002000

TRT INetwork construction elapsed time: 0:00:00.002000

2023-03-26 19:55:23.835

Build TRT engine elapsed time: 0:00:04.250712

Build TRT engine elapsed time: 0:00:04.250712

Lowering submodule _run_on_acc_4 elapsed time 0:00:04.281701

Lowering submodule _run_on_acc_4 elapsed time 0:00:04.281701

Now lowering submodule _run_on_acc_6

Now lowering submodule _run_on_acc_6

split_name=_run_on_acc_6, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_6, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.001000

TRT INetwork construction elapsed time: 0:00:00.001000

2023-03-26 19:55:28.104

Build TRT engine elapsed time: 0:00:04.225747

Build TRT engine elapsed time: 0:00:04.225747

Lowering submodule _run_on_acc_6 elapsed time 0:00:04.254747

Lowering submodule _run_on_acc_6 elapsed time 0:00:04.254747

Now lowering submodule _run_on_acc_8

Now lowering submodule _run_on_acc_8

split_name=_run_on_acc_8, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_8, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.003000

TRT INetwork construction elapsed time: 0:00:00.003000

2023-03-26 19:55:33.197

Build TRT engine elapsed time: 0:00:05.048240

Build TRT engine elapsed time: 0:00:05.048240

Lowering submodule _run_on_acc_8 elapsed time 0:00:05.080299

Lowering submodule _run_on_acc_8 elapsed time 0:00:05.080299

Now lowering submodule _run_on_acc_10

Now lowering submodule _run_on_acc_10

split_name=_run_on_acc_10, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_10, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.001349

TRT INetwork construction elapsed time: 0:00:00.001349

Build TRT engine elapsed time: 0:00:01.816100

Build TRT engine elapsed time: 0:00:01.816100

Lowering submodule _run_on_acc_10 elapsed time 0:00:01.845110

Lowering submodule _run_on_acc_10 elapsed time 0:00:01.845110

Now lowering submodule _run_on_acc_12

Now lowering submodule _run_on_acc_12

split_name=_run_on_acc_12, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_12, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.002249

TRT INetwork construction elapsed time: 0:00:00.002249

Build TRT engine elapsed time: 0:00:01.737383

Build TRT engine elapsed time: 0:00:01.737383

Lowering submodule _run_on_acc_12 elapsed time 0:00:01.770741

Lowering submodule _run_on_acc_12 elapsed time 0:00:01.770741

Now lowering submodule _run_on_acc_14

Now lowering submodule _run_on_acc_14

split_name=_run_on_acc_14, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_14, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.001000

TRT INetwork construction elapsed time: 0:00:00.001000

Build TRT engine elapsed time: 0:00:01.776626

Build TRT engine elapsed time: 0:00:01.776626

Lowering submodule _run_on_acc_14 elapsed time 0:00:01.807626

Lowering submodule _run_on_acc_14 elapsed time 0:00:01.807626

Now lowering submodule _run_on_acc_16

Now lowering submodule _run_on_acc_16

split_name=_run_on_acc_16, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_16, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.001001

TRT INetwork construction elapsed time: 0:00:00.001001

Build TRT engine elapsed time: 0:00:01.797982

Build TRT engine elapsed time: 0:00:01.797982

Lowering submodule _run_on_acc_16 elapsed time 0:00:01.827555

Lowering submodule _run_on_acc_16 elapsed time 0:00:01.827555

Now lowering submodule _run_on_acc_18

Now lowering submodule _run_on_acc_18

split_name=_run_on_acc_18, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 14080]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_18, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 14080]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-26 19:55:40.535

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-26 19:55:40.542

TRT INetwork construction elapsed time: 0:00:00.007002

TRT INetwork construction elapsed time: 0:00:00.007002

2023-03-26 19:56:02.334

Build TRT engine elapsed time: 0:00:21.785003

Build TRT engine elapsed time: 0:00:21.785003

Lowering submodule _run_on_acc_18 elapsed time 0:00:21.820629

Lowering submodule _run_on_acc_18 elapsed time 0:00:21.820629

Now lowering submodule _run_on_acc_20

Now lowering submodule _run_on_acc_20

split_name=_run_on_acc_20, input_specs=[InputTensorSpec(shape=torch.Size([1, 88, 160, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 88, 160, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_20, input_specs=[InputTensorSpec(shape=torch.Size([1, 88, 160, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 88, 160, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_144 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_144 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_145 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_145 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_146 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_146 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_147 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_147 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.032007

TRT INetwork construction elapsed time: 0:00:00.032007

Build TRT engine elapsed time: 0:00:00.427231

Build TRT engine elapsed time: 0:00:00.427231

Lowering submodule _run_on_acc_20 elapsed time 0:00:00.488966

Lowering submodule _run_on_acc_20 elapsed time 0:00:00.488966

Now lowering submodule _run_on_acc_22

Now lowering submodule _run_on_acc_22

split_name=_run_on_acc_22, input_specs=[InputTensorSpec(shape=torch.Size([220, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([220, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 220, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_22, input_specs=[InputTensorSpec(shape=torch.Size([220, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([220, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 220, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_148 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_148 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_149 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_149 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.019005

TRT INetwork construction elapsed time: 0:00:00.019005

2023-03-26 19:56:08.464

Failed to evaluate the script:
Python exception:

Traceback (most recent call last):
File "src\cython\vapoursynth.pyx", line 2866, in vapoursynth._vpy_evaluate
File "src\cython\vapoursynth.pyx", line 2867, in vapoursynth._vpy_evaluate
File "J:\tmp\tempPreviewVapoursynthFile19_54_46_507.vpy", line 36, in
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch\autograd\grad_mode.py", line 27, in decorate_context
return func(*args, **kwargs)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\vsfemasr\__init__.py", line 171, in femasr
module = lowerer(
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\lower.py", line 323, in __call__
return do_lower(module, inputs)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\passes\pass_utils.py", line 117, in pass_with_validation
processed_module = pass_(module, input, *args, **kwargs)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\lower.py", line 320, in do_lower
lower_result = pm(module)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch\fx\passes\pass_manager.py", line 240, in __call__
out = _pass(out)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch\fx\passes\pass_manager.py", line 240, in __call__
out = _pass(out)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\passes\lower_pass_manager_builder.py", line 167, in lower_func
lowered_module = self._lower_func(
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\lower.py", line 180, in lower_pass
interp_res: TRTInterpreterResult = interpreter(mod, input, module_name)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\lower.py", line 132, in __call__
interp_result: TRTInterpreterResult = interpreter.run(
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\fx2trt.py", line 252, in run
assert engine
AssertionError

2023-03-26 19:56:08.622

[VSE Server]: incoming connection

[VSE Server]: ConnectedState

[VSE Server]: socket is ready to be read

[VSE Server]: connection open: true

[VSE Server]: connection readable: true

[VSE Server] - Message received: changeTo ### J:\tmp\tempPreviewVapoursynthFile19_55_25_204.vpy ### off#0#0#0#0

2023-03-26 19:56:09.052

Skip rewriting leaf module

Skip rewriting leaf module

Skip rewriting leaf module

Skip rewriting leaf module

2023-03-26 19:56:14.029

== Log pass before/after graph to C:\Users\Selur\AppData\Local\Temp\tmpreq3dqbl, before/after are the same = False

== Log pass before/after graph to C:\Users\Selur\AppData\Local\Temp\tmpreq3dqbl, before/after are the same = False

== Log pass before/after graph to C:\Users\Selur\AppData\Local\Temp\tmpgrxuz6hl, before/after are the same = True

== Log pass before/after graph to C:\Users\Selur\AppData\Local\Temp\tmpgrxuz6hl, before/after are the same = True

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_319

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_319

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_326

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_326

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_329

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_329

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_331

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_331

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_332

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_332

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_339

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_339

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_342

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_342

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_344

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_344

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_345

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_345

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_352

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_352

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_355

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_355

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_357

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_357

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_358

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_358

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_365

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_365

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_368

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_368

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_370

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_370

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_371

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_371

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_378

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_378

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_381

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_381

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_383

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_383

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_384

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_384

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_391

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_391

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_394

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_394

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_396

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_396

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_397

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_397

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_398

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_398

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_405

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_405

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_408

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_408

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_410

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_410

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_411

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_411

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_418

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_418

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_421

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_421

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_423

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_423

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_424

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_424

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_431

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_431

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_434

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_434

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_436

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_436

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_437

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_437

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_444

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_444

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_447

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_447

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_449

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_449

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_450

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_450

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_457

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_457

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_460

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_460

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_462

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_462

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_463

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_463

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_470

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_470

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_473

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_473

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_475

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_475

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_476

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_476

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_477

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_477

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_484

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_484

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_487

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_487

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_489

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_489

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_490

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_490

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_497

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_497

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_500

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_500

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_502

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_502

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_503

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_503

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_510

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_510

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_513

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_513

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_515

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_515

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_516

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_516

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_523

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_523

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_526

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_526

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_528

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_528

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_529

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_529

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_536

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_536

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_539

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_539

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_541

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_541

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_542

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_542

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_549

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_549

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_552

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_552

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_554

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_554

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_555

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_555

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_556

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_556

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_563

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_563

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_566

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_566

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_568

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_568

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_569

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_569

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_576

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_576

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_579

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_579

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_581

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_581

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_582

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_582

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_589

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_589

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_592

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_592

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_594

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_594

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_595

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_595

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_602

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_602

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_605

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_605

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_607

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_607

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_608

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_608

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_615

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_615

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_618

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_618

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_620

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_620

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_621

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_621

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_628

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_628

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_631

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_631

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_633

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_633

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_634

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_634

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_635

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_635

Now lowering submodule _run_on_acc_0

Now lowering submodule _run_on_acc_0

split_name=_run_on_acc_0, input_specs=[InputTensorSpec(shape=torch.Size([1, 3, 352, 640]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_0, input_specs=[InputTensorSpec(shape=torch.Size([1, 3, 352, 640]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.416462

TRT INetwork construction elapsed time: 0:00:00.416462

2023-03-26 19:56:20.745

Build TRT engine elapsed time: 0:00:02.908010

Build TRT engine elapsed time: 0:00:02.908010

Lowering submodule _run_on_acc_0 elapsed time 0:00:03.358515

Lowering submodule _run_on_acc_0 elapsed time 0:00:03.358515

Now lowering submodule _run_on_acc_2

Now lowering submodule _run_on_acc_2

split_name=_run_on_acc_2, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_2, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.001000

TRT INetwork construction elapsed time: 0:00:00.001000

2023-03-26 19:56:25.139

Build TRT engine elapsed time: 0:00:04.345430

Build TRT engine elapsed time: 0:00:04.345430

Lowering submodule _run_on_acc_2 elapsed time 0:00:04.378391

Lowering submodule _run_on_acc_2 elapsed time 0:00:04.378391

Now lowering submodule _run_on_acc_4

Now lowering submodule _run_on_acc_4

split_name=_run_on_acc_4, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_4, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.001000

TRT INetwork construction elapsed time: 0:00:00.001000

2023-03-26 19:56:29.557

Build TRT engine elapsed time: 0:00:04.366875

Build TRT engine elapsed time: 0:00:04.366875

Lowering submodule _run_on_acc_4 elapsed time 0:00:04.403853

Lowering submodule _run_on_acc_4 elapsed time 0:00:04.403853

Now lowering submodule _run_on_acc_6

Now lowering submodule _run_on_acc_6

split_name=_run_on_acc_6, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_6, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.002001

TRT INetwork construction elapsed time: 0:00:00.002001

2023-03-26 19:56:33.951

Build TRT engine elapsed time: 0:00:04.345732

Build TRT engine elapsed time: 0:00:04.345732

Lowering submodule _run_on_acc_6 elapsed time 0:00:04.378103

Lowering submodule _run_on_acc_6 elapsed time 0:00:04.378103

Now lowering submodule _run_on_acc_8

Now lowering submodule _run_on_acc_8

split_name=_run_on_acc_8, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_8, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.003001

TRT INetwork construction elapsed time: 0:00:00.003001

2023-03-26 19:56:39.176

Build TRT engine elapsed time: 0:00:05.176863

Build TRT engine elapsed time: 0:00:05.176863

Lowering submodule _run_on_acc_8 elapsed time 0:00:05.211868

Lowering submodule _run_on_acc_8 elapsed time 0:00:05.211868

Now lowering submodule _run_on_acc_10

Now lowering submodule _run_on_acc_10

split_name=_run_on_acc_10, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_10, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.001001

TRT INetwork construction elapsed time: 0:00:00.001001

Build TRT engine elapsed time: 0:00:01.780700

Build TRT engine elapsed time: 0:00:01.780700

Lowering submodule _run_on_acc_10 elapsed time 0:00:01.815286

Lowering submodule _run_on_acc_10 elapsed time 0:00:01.815286

Now lowering submodule _run_on_acc_12

Now lowering submodule _run_on_acc_12

split_name=_run_on_acc_12, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_12, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.002000

TRT INetwork construction elapsed time: 0:00:00.002000

Build TRT engine elapsed time: 0:00:01.784465

Build TRT engine elapsed time: 0:00:01.784465

Lowering submodule _run_on_acc_12 elapsed time 0:00:01.820483

Lowering submodule _run_on_acc_12 elapsed time 0:00:01.820483

Now lowering submodule _run_on_acc_14

Now lowering submodule _run_on_acc_14

split_name=_run_on_acc_14, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_14, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.002000

TRT INetwork construction elapsed time: 0:00:00.002000

Build TRT engine elapsed time: 0:00:01.792631

Build TRT engine elapsed time: 0:00:01.792631

Lowering submodule _run_on_acc_14 elapsed time 0:00:01.824090

Lowering submodule _run_on_acc_14 elapsed time 0:00:01.824090

Now lowering submodule _run_on_acc_16

Now lowering submodule _run_on_acc_16

split_name=_run_on_acc_16, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_16, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.001000

TRT INetwork construction elapsed time: 0:00:00.001000

Build TRT engine elapsed time: 0:00:01.791811

Build TRT engine elapsed time: 0:00:01.791811

Lowering submodule _run_on_acc_16 elapsed time 0:00:01.822825

Lowering submodule _run_on_acc_16 elapsed time 0:00:01.822825

Now lowering submodule _run_on_acc_18

Now lowering submodule _run_on_acc_18

split_name=_run_on_acc_18, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 14080]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_18, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 14080]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-26 19:56:46.553

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-26 19:56:46.561

TRT INetwork construction elapsed time: 0:00:00.008002

TRT INetwork construction elapsed time: 0:00:00.008002

2023-03-26 19:56:52.869

Build TRT engine elapsed time: 0:00:06.301220

Build TRT engine elapsed time: 0:00:06.301220

Lowering submodule _run_on_acc_18 elapsed time 0:00:06.339860

Lowering submodule _run_on_acc_18 elapsed time 0:00:06.339860

Now lowering submodule _run_on_acc_20

Now lowering submodule _run_on_acc_20

split_name=_run_on_acc_20, input_specs=[InputTensorSpec(shape=torch.Size([1, 88, 160, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 88, 160, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_20, input_specs=[InputTensorSpec(shape=torch.Size([1, 88, 160, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 88, 160, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.004995

TRT INetwork construction elapsed time: 0:00:00.004995

Build TRT engine elapsed time: 0:00:00.415419

Build TRT engine elapsed time: 0:00:00.415419

Lowering submodule _run_on_acc_20 elapsed time 0:00:00.451659

Lowering submodule _run_on_acc_20 elapsed time 0:00:00.451659

Now lowering submodule _run_on_acc_22

Now lowering submodule _run_on_acc_22

split_name=_run_on_acc_22, input_specs=[InputTensorSpec(shape=torch.Size([220, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([220, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 220, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_22, input_specs=[InputTensorSpec(shape=torch.Size([220, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([220, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 220, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.005986

TRT INetwork construction elapsed time: 0:00:00.005986

2023-03-26 19:57:08.059

Failed to evaluate the script:
Python exception:

Traceback (most recent call last):
File "src\cython\vapoursynth.pyx", line 2866, in vapoursynth._vpy_evaluate
File "src\cython\vapoursynth.pyx", line 2867, in vapoursynth._vpy_evaluate
File "J:\tmp\tempPreviewVapoursynthFile19_55_25_204.vpy", line 36, in
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch\autograd\grad_mode.py", line 27, in decorate_context
return func(*args, **kwargs)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\vsfemasr\__init__.py", line 171, in femasr
module = lowerer(
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\lower.py", line 323, in __call__
return do_lower(module, inputs)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\passes\pass_utils.py", line 117, in pass_with_validation
processed_module = pass_(module, input, *args, **kwargs)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\lower.py", line 320, in do_lower
lower_result = pm(module)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch\fx\passes\pass_manager.py", line 240, in __call__
out = _pass(out)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch\fx\passes\pass_manager.py", line 240, in __call__
out = _pass(out)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\passes\lower_pass_manager_builder.py", line 167, in lower_func
lowered_module = self._lower_func(
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\lower.py", line 180, in lower_pass
interp_res: TRTInterpreterResult = interpreter(mod, input, module_name)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\lower.py", line 132, in __call__
interp_result: TRTInterpreterResult = interpreter.run(
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\fx2trt.py", line 252, in run
assert engine
AssertionError

2023-03-26 19:57:08.195

[VSE Server]: socket is ready to be read

[VSE Server]: connection open: true

[VSE Server]: connection readable: true

[VSE Server] - Message received: changeTo ### J:\tmp\tempPreviewVapoursynthFile19_56_10_330.vpy ### off#0#0#0#0

2023-03-26 19:57:08.704

Skip rewriting leaf module

Skip rewriting leaf module

Skip rewriting leaf module

Skip rewriting leaf module

2023-03-26 19:57:13.600

== Log pass before/after graph to C:\Users\Selur\AppData\Local\Temp\tmpa_f7h_70, before/after are the same = False

== Log pass before/after graph to C:\Users\Selur\AppData\Local\Temp\tmpa_f7h_70, before/after are the same = False

== Log pass before/after graph to C:\Users\Selur\AppData\Local\Temp\tmp6nkjkumw, before/after are the same = True

== Log pass before/after graph to C:\Users\Selur\AppData\Local\Temp\tmp6nkjkumw, before/after are the same = True

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_319

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_319

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_326

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_326

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_329

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_329

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_331

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_331

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_332

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_332

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_339

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_339

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_342

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_342

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_344

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_344

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_345

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_345

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_352

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_352

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_355

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_355

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_357

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_357

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_358

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_358

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_365

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_365

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_368

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_368

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_370

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_370

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_371

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_371

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_378

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_378

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_381

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_381

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_383

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_383

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_384

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_384

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_391

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_391

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_394

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_394

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_396

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_396

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_397

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_397

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_398

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_398

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_405

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_405

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_408

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_408

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_410

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_410

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_411

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_411

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_418

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_418

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_421

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_421

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_423

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_423

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_424

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_424

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_431

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_431

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_434

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_434

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_436

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_436

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_437

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_437

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_444

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_444

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_447

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_447

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_449

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_449

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_450

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_450

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_457

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_457

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_460

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_460

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_462

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_462

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_463

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_463

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_470

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_470

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_473

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_473

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_475

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_475

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_476

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_476

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_477

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_477

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_484

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_484

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_487

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_487

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_489

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_489

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_490

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_490

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_497

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_497

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_500

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_500

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_502

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_502

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_503

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_503

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_510

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_510

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_513

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_513

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_515

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_515

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_516

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_516

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_523

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_523

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_526

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_526

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_528

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_528

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_529

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_529

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_536

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_536

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_539

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_539

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_541

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_541

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_542

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_542

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_549

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_549

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_552

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_552

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_554

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_554

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_555

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_555

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_556

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_556

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_563

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_563

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_566

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_566

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_568

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_568

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_569

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_569

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_576

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_576

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_579

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_579

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_581

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_581

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_582

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_582

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_589

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_589

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_592

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_592

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_594

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_594

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_595

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_595

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_602

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_602

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_605

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_605

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_607

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_607

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_608

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_608

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_615

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_615

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_618

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_618

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_620

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_620

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_621

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_621

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_628

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_628

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_631

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_631

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_633

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_633

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_634

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_634

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_635

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_635

Now lowering submodule _run_on_acc_0

Now lowering submodule _run_on_acc_0

split_name=_run_on_acc_0, input_specs=[InputTensorSpec(shape=torch.Size([1, 3, 352, 640]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_0, input_specs=[InputTensorSpec(shape=torch.Size([1, 3, 352, 640]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.283551

TRT INetwork construction elapsed time: 0:00:00.283551

2023-03-26 19:57:25.432

Build TRT engine elapsed time: 0:00:02.833113

Build TRT engine elapsed time: 0:00:02.833113

Lowering submodule _run_on_acc_0 elapsed time 0:00:03.308137

Lowering submodule _run_on_acc_0 elapsed time 0:00:03.308137

Now lowering submodule _run_on_acc_2

Now lowering submodule _run_on_acc_2

split_name=_run_on_acc_2, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_2, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.002000

TRT INetwork construction elapsed time: 0:00:00.002000

2023-03-26 19:57:29.901

Build TRT engine elapsed time: 0:00:04.184779

Build TRT engine elapsed time: 0:00:04.184779

Lowering submodule _run_on_acc_2 elapsed time 0:00:04.375815

Lowering submodule _run_on_acc_2 elapsed time 0:00:04.375815

Now lowering submodule _run_on_acc_4

Now lowering submodule _run_on_acc_4

split_name=_run_on_acc_4, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_4, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.001000

TRT INetwork construction elapsed time: 0:00:00.001000

2023-03-26 19:57:34.431

Build TRT engine elapsed time: 0:00:04.242936

Build TRT engine elapsed time: 0:00:04.242936

Lowering submodule _run_on_acc_4 elapsed time 0:00:04.439479

Lowering submodule _run_on_acc_4 elapsed time 0:00:04.439479

Now lowering submodule _run_on_acc_6

Now lowering submodule _run_on_acc_6

split_name=_run_on_acc_6, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_6, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.001000

TRT INetwork construction elapsed time: 0:00:00.001000

2023-03-26 19:57:38.908

Build TRT engine elapsed time: 0:00:04.185875

Build TRT engine elapsed time: 0:00:04.185875

Lowering submodule _run_on_acc_6 elapsed time 0:00:04.380859

Lowering submodule _run_on_acc_6 elapsed time 0:00:04.380859

Now lowering submodule _run_on_acc_8

Now lowering submodule _run_on_acc_8

split_name=_run_on_acc_8, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_8, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 176, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.003000

TRT INetwork construction elapsed time: 0:00:00.003000

2023-03-26 19:57:44.242

Build TRT engine elapsed time: 0:00:05.038204

Build TRT engine elapsed time: 0:00:05.038204

Lowering submodule _run_on_acc_8 elapsed time 0:00:05.239721

Lowering submodule _run_on_acc_8 elapsed time 0:00:05.239721

Now lowering submodule _run_on_acc_10

Now lowering submodule _run_on_acc_10

split_name=_run_on_acc_10, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_10, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.001001

TRT INetwork construction elapsed time: 0:00:00.001001

Build TRT engine elapsed time: 0:00:01.729166

Build TRT engine elapsed time: 0:00:01.729166

Lowering submodule _run_on_acc_10 elapsed time 0:00:01.926857

Lowering submodule _run_on_acc_10 elapsed time 0:00:01.926857

Now lowering submodule _run_on_acc_12

Now lowering submodule _run_on_acc_12

split_name=_run_on_acc_12, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_12, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.002000

TRT INetwork construction elapsed time: 0:00:00.002000

Build TRT engine elapsed time: 0:00:01.723373

Build TRT engine elapsed time: 0:00:01.723373

Lowering submodule _run_on_acc_12 elapsed time 0:00:01.923711

Lowering submodule _run_on_acc_12 elapsed time 0:00:01.923711

Now lowering submodule _run_on_acc_14

Now lowering submodule _run_on_acc_14

split_name=_run_on_acc_14, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_14, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.002000

TRT INetwork construction elapsed time: 0:00:00.002000

Build TRT engine elapsed time: 0:00:01.731051

Build TRT engine elapsed time: 0:00:01.731051

Lowering submodule _run_on_acc_14 elapsed time 0:00:01.931001

Lowering submodule _run_on_acc_14 elapsed time 0:00:01.931001

Now lowering submodule _run_on_acc_16

Now lowering submodule _run_on_acc_16

split_name=_run_on_acc_16, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_16, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 88, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.000999

TRT INetwork construction elapsed time: 0:00:00.000999

Build TRT engine elapsed time: 0:00:01.724712

Build TRT engine elapsed time: 0:00:01.724712

Lowering submodule _run_on_acc_16 elapsed time 0:00:01.928761

Lowering submodule _run_on_acc_16 elapsed time 0:00:01.928761

Now lowering submodule _run_on_acc_18

Now lowering submodule _run_on_acc_18

split_name=_run_on_acc_18, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 14080]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_18, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 14080]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-26 19:57:52.589

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-26 19:57:52.640

TRT INetwork construction elapsed time: 0:00:00.052510

TRT INetwork construction elapsed time: 0:00:00.052510

2023-03-26 19:57:58.768

Build TRT engine elapsed time: 0:00:06.077423

Build TRT engine elapsed time: 0:00:06.077423

Lowering submodule _run_on_acc_18 elapsed time 0:00:06.330382

Lowering submodule _run_on_acc_18 elapsed time 0:00:06.330382

Now lowering submodule _run_on_acc_20

Now lowering submodule _run_on_acc_20

split_name=_run_on_acc_20, input_specs=[InputTensorSpec(shape=torch.Size([1, 88, 160, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 88, 160, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_20, input_specs=[InputTensorSpec(shape=torch.Size([1, 88, 160, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 88, 160, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.006000

TRT INetwork construction elapsed time: 0:00:00.006000

Build TRT engine elapsed time: 0:00:00.414580

Build TRT engine elapsed time: 0:00:00.414580

Lowering submodule _run_on_acc_20 elapsed time 0:00:00.638044

Lowering submodule _run_on_acc_20 elapsed time 0:00:00.638044

Now lowering submodule _run_on_acc_22

Now lowering submodule _run_on_acc_22

split_name=_run_on_acc_22, input_specs=[InputTensorSpec(shape=torch.Size([220, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([220, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 220, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_22, input_specs=[InputTensorSpec(shape=torch.Size([220, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([220, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 220, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.007001

TRT INetwork construction elapsed time: 0:00:00.007001

2023-03-26 19:58:04.459

Failed to evaluate the script:
Python exception:

Traceback (most recent call last):
File "src\cython\vapoursynth.pyx", line 2866, in vapoursynth._vpy_evaluate
File "src\cython\vapoursynth.pyx", line 2867, in vapoursynth._vpy_evaluate
File "J:\tmp\tempPreviewVapoursynthFile19_56_10_330.vpy", line 36, in
clip = FeMaSR(clip=clip, device_index=0, trt=True, trt_cache_path=r"J:\tmp") # 1280x704
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch\autograd\grad_mode.py", line 27, in decorate_context
return func(*args, **kwargs)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\vsfemasr\__init__.py", line 171, in femasr
module = lowerer(
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\lower.py", line 323, in __call__
return do_lower(module, inputs)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\passes\pass_utils.py", line 117, in pass_with_validation
processed_module = pass_(module, input, *args, **kwargs)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\lower.py", line 320, in do_lower
lower_result = pm(module)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch\fx\passes\pass_manager.py", line 240, in __call__
out = _pass(out)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch\fx\passes\pass_manager.py", line 240, in __call__
out = _pass(out)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\passes\lower_pass_manager_builder.py", line 167, in lower_func
lowered_module = self._lower_func(
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\lower.py", line 180, in lower_pass
interp_res: TRTInterpreterResult = interpreter(mod, input, module_name)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\lower.py", line 132, in __call__
interp_result: TRTInterpreterResult = interpreter.run(
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\fx2trt.py", line 252, in run
assert engine
AssertionError
Selur commented 1 year ago

Using lower resolution:

# Imports
import vapoursynth as vs
# getting Vapoursynth core
core = vs.core
import site
import os
import ctypes
# Adding torch dependencies to PATH
path = site.getsitepackages()[0]+'/torch_dependencies/bin/'
ctypes.windll.kernel32.SetDllDirectoryW(path)
path = path.replace('\\', '/')
os.environ["PATH"] = path + os.pathsep + os.environ["PATH"]
os.environ["CUDA_MODULE_LOADING"] = "LAZY"
# Loading Plugins
core.std.LoadPlugin(path="i:/Hybrid/64bit/vsfilters/Support/fmtconv.dll")
core.std.LoadPlugin(path="i:/Hybrid/64bit/vsfilters/SourceFilter/LSmashSource/vslsmashsource.dll")
# source: 'G:\TestClips&Co\test.avi'
# current color space: YUV420P8, bit depth: 8, resolution: 640x352, fps: 25, color matrix: 470bg, yuv luminance scale: limited, scanorder: progressive
# Loading G:\TestClips&Co\test.avi using LWLibavSource
clip = core.lsmas.LWLibavSource(source="G:/TestClips&Co/test.avi", format="YUV420P8", stream_index=0, cache=0, prefer_hw=0)
# Setting detected color matrix (470bg).
clip = core.std.SetFrameProps(clip, _Matrix=5)
# Setting color transfer info (470bg), when it is not set
clip = clip if not core.text.FrameProps(clip,'_Transfer') else core.std.SetFrameProps(clip, _Transfer=5)
# Setting color primaries info (), when it is not set
clip = clip if not core.text.FrameProps(clip,'_Primaries') else core.std.SetFrameProps(clip, _Primaries=5)
# Setting color range to TV (limited) range.
clip = core.std.SetFrameProp(clip=clip, prop="_ColorRange", intval=1)
# making sure frame rate is set to 25
clip = core.std.AssumeFPS(clip=clip, fpsnum=25, fpsden=1)
clip = core.std.SetFrameProp(clip=clip, prop="_FieldBased", intval=0) # progressive
# adjusting resolution before resizing 
clip = core.fmtc.resample(clip=clip, w=320, h=176, kernel="lanczos", interlaced=False, interlacedd=False)# before YUV420P8 after YUV420P16
from vsfemasr import femasr as FeMaSR
# adjusting color space from YUV420P16 to RGBH for VsFeMaSR
clip = core.resize.Bicubic(clip=clip, format=vs.RGBH, matrix_in_s="470bg", range_s="limited")
# resizing using FeMaSR
clip = FeMaSR(clip=clip, device_index=0, trt=True, trt_cache_path=r"J:\tmp") # 640x352
# adjusting output color from: RGBH to YUV420P10 for x265Model
clip = core.resize.Bicubic(clip=clip, format=vs.YUV420P10, matrix_s="470bg", range_s="limited", dither_type="error_diffusion")
# set output frame rate to 25fps (progressive)
clip = core.std.AssumeFPS(clip=clip, fpsnum=25, fpsden=1)
# Output
clip.set_output()

doesn't help either:


2023-03-27 05:19:43.377

Skip rewriting leaf module

Skip rewriting leaf module

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\tracer\acc_tracer\acc_tracer.py:584: UserWarning: acc_tracer does not support currently support models for training. Calling eval on model before tracing.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\tracer\acc_tracer\acc_tracer.py:584: UserWarning: acc_tracer does not support currently support models for training. Calling eval on model before tracing.
warnings.warn(

Skip rewriting leaf module

Skip rewriting leaf module

== Log pass before/after graph to C:\Users\Selur\AppData\Local\Temp\tmpjmqoifyb, before/after are the same = False

== Log pass before/after graph to C:\Users\Selur\AppData\Local\Temp\tmpjmqoifyb, before/after are the same = False

== Log pass before/after graph to C:\Users\Selur\AppData\Local\Temp\tmpu2_5tp_s, before/after are the same = True

== Log pass before/after graph to C:\Users\Selur\AppData\Local\Temp\tmpu2_5tp_s, before/after are the same = True

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_319

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_319

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_326

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_326

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_329

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_329

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_331

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_331

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_332

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_332

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_339

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_339

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_342

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_342

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_344

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_344

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_345

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_345

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_352

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_352

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_355

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_355

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_357

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_357

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_358

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_358

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_365

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_365

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_368

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_368

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_370

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_370

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_371

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_371

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_378

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_378

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_381

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_381

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_383

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_383

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_384

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_384

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_391

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_391

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_394

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_394

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_396

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_396

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_397

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_397

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_398

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_398

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_405

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_405

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_408

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_408

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_410

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_410

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_411

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_411

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_418

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_418

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_421

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_421

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_423

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_423

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_424

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_424

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_431

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_431

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_434

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_434

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_436

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_436

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_437

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_437

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_444

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_444

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_447

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_447

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_449

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_449

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_450

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_450

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_457

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_457

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_460

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_460

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_462

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_462

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_463

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_463

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_470

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_470

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_473

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_473

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_475

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_475

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_476

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_476

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_477

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_477

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_484

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_484

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_487

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_487

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_489

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_489

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_490

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_490

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_497

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_497

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_500

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_500

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_502

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_502

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_503

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_503

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_510

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_510

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_513

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_513

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_515

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_515

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_516

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_516

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_523

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_523

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_526

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_526

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_528

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_528

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_529

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_529

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_536

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_536

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_539

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_539

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_541

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_541

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_542

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_542

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_549

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_549

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_552

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_552

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_554

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_554

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_555

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_555

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_556

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_556

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_563

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_563

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_566

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_566

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_568

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_568

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_569

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_569

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_576

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_576

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_579

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_579

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_581

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_581

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_582

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_582

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_589

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_589

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_592

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_592

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_594

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_594

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_595

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_595

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_602

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_602

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_605

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_605

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_607

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_607

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_608

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_608

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_615

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_615

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_618

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_618

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_620

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_620

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_621

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_621

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_628

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_628

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_631

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_631

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_633

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_633

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_634

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_634

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_635

: Found bad pattern: y.reshape((x, ...)). Reshape node: reshape_635

Now lowering submodule _run_on_acc_0

Now lowering submodule _run_on_acc_0

split_name=_run_on_acc_0, input_specs=[InputTensorSpec(shape=torch.Size([1, 3, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_0, input_specs=[InputTensorSpec(shape=torch.Size([1, 3, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:19:52.711

TRT INetwork construction elapsed time: 0:00:00.007916

TRT INetwork construction elapsed time: 0:00:00.007916

Build TRT engine elapsed time: 0:00:01.467011

Build TRT engine elapsed time: 0:00:01.467011

Lowering submodule _run_on_acc_0 elapsed time 0:00:06.490400

Lowering submodule _run_on_acc_0 elapsed time 0:00:06.490400

Now lowering submodule _run_on_acc_2

Now lowering submodule _run_on_acc_2

split_name=_run_on_acc_2, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_2, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.002721

TRT INetwork construction elapsed time: 0:00:00.002721

2023-03-27 05:20:13.327

Build TRT engine elapsed time: 0:00:19.097724

Build TRT engine elapsed time: 0:00:19.097724

Lowering submodule _run_on_acc_2 elapsed time 0:00:19.127441

Lowering submodule _run_on_acc_2 elapsed time 0:00:19.127441

Now lowering submodule _run_on_acc_4

Now lowering submodule _run_on_acc_4

split_name=_run_on_acc_4, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_4, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.000997

TRT INetwork construction elapsed time: 0:00:00.000997

Build TRT engine elapsed time: 0:00:01.871680

Build TRT engine elapsed time: 0:00:01.871680

Lowering submodule _run_on_acc_4 elapsed time 0:00:01.900656

Lowering submodule _run_on_acc_4 elapsed time 0:00:01.900656

Now lowering submodule _run_on_acc_6

Now lowering submodule _run_on_acc_6

split_name=_run_on_acc_6, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_6, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.000997

TRT INetwork construction elapsed time: 0:00:00.000997

Build TRT engine elapsed time: 0:00:01.830022

Build TRT engine elapsed time: 0:00:01.830022

Lowering submodule _run_on_acc_6 elapsed time 0:00:01.859147

Lowering submodule _run_on_acc_6 elapsed time 0:00:01.859147

Now lowering submodule _run_on_acc_8

Now lowering submodule _run_on_acc_8

split_name=_run_on_acc_8, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_8, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.003000

TRT INetwork construction elapsed time: 0:00:00.003000

2023-03-27 05:20:19.471

Build TRT engine elapsed time: 0:00:02.317220

Build TRT engine elapsed time: 0:00:02.317220

Lowering submodule _run_on_acc_8 elapsed time 0:00:02.347223

Lowering submodule _run_on_acc_8 elapsed time 0:00:02.347223

Now lowering submodule _run_on_acc_10

Now lowering submodule _run_on_acc_10

split_name=_run_on_acc_10, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_10, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.000999

TRT INetwork construction elapsed time: 0:00:00.000999

Build TRT engine elapsed time: 0:00:01.354626

Build TRT engine elapsed time: 0:00:01.354626

Lowering submodule _run_on_acc_10 elapsed time 0:00:01.382135

Lowering submodule _run_on_acc_10 elapsed time 0:00:01.382135

Now lowering submodule _run_on_acc_12

Now lowering submodule _run_on_acc_12

split_name=_run_on_acc_12, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_12, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.001000

TRT INetwork construction elapsed time: 0:00:00.001000

Build TRT engine elapsed time: 0:00:01.378106

Build TRT engine elapsed time: 0:00:01.378106

Lowering submodule _run_on_acc_12 elapsed time 0:00:01.407108

Lowering submodule _run_on_acc_12 elapsed time 0:00:01.407108

Now lowering submodule _run_on_acc_14

Now lowering submodule _run_on_acc_14

split_name=_run_on_acc_14, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_14, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.000999

TRT INetwork construction elapsed time: 0:00:00.000999

Build TRT engine elapsed time: 0:00:01.380241

Build TRT engine elapsed time: 0:00:01.380241

Lowering submodule _run_on_acc_14 elapsed time 0:00:01.410777

Lowering submodule _run_on_acc_14 elapsed time 0:00:01.410777

Now lowering submodule _run_on_acc_16

Now lowering submodule _run_on_acc_16

split_name=_run_on_acc_16, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_16, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.003001

TRT INetwork construction elapsed time: 0:00:00.003001

Build TRT engine elapsed time: 0:00:01.396178

Build TRT engine elapsed time: 0:00:01.396178

Lowering submodule _run_on_acc_16 elapsed time 0:00:01.426690

Lowering submodule _run_on_acc_16 elapsed time 0:00:01.426690

Now lowering submodule _run_on_acc_18

Now lowering submodule _run_on_acc_18

split_name=_run_on_acc_18, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 3840]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_18, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 3840]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:20:25.185

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:20:25.191

TRT INetwork construction elapsed time: 0:00:00.010184

TRT INetwork construction elapsed time: 0:00:00.010184

2023-03-27 05:20:43.247

Build TRT engine elapsed time: 0:00:18.049577

Build TRT engine elapsed time: 0:00:18.049577

Lowering submodule _run_on_acc_18 elapsed time 0:00:18.085758

Lowering submodule _run_on_acc_18 elapsed time 0:00:18.085758

Now lowering submodule _run_on_acc_20

Now lowering submodule _run_on_acc_20

split_name=_run_on_acc_20, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_20, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_144 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_144 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_145 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_145 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_146 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_146 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_147 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_147 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.033131

TRT INetwork construction elapsed time: 0:00:00.033131

Build TRT engine elapsed time: 0:00:00.472221

Build TRT engine elapsed time: 0:00:00.472221

Lowering submodule _run_on_acc_20 elapsed time 0:00:00.534362

Lowering submodule _run_on_acc_20 elapsed time 0:00:00.534362

Now lowering submodule _run_on_acc_22

Now lowering submodule _run_on_acc_22

split_name=_run_on_acc_22, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_22, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_148 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_148 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_149 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_149 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.021358

TRT INetwork construction elapsed time: 0:00:00.021358

2023-03-27 05:20:48.278

Build TRT engine elapsed time: 0:00:04.422826

Build TRT engine elapsed time: 0:00:04.422826

Lowering submodule _run_on_acc_22 elapsed time 0:00:04.474202

Lowering submodule _run_on_acc_22 elapsed time 0:00:04.474202

Now lowering submodule _run_on_acc_24

Now lowering submodule _run_on_acc_24

split_name=_run_on_acc_24, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_24, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:20:48.317

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:20:48.335

TRT INetwork construction elapsed time: 0:00:00.019704

TRT INetwork construction elapsed time: 0:00:00.019704

2023-03-27 05:20:58.364

Build TRT engine elapsed time: 0:00:10.021728

Build TRT engine elapsed time: 0:00:10.021728

Lowering submodule _run_on_acc_24 elapsed time 0:00:10.068910

Lowering submodule _run_on_acc_24 elapsed time 0:00:10.068910

Now lowering submodule _run_on_acc_26

Now lowering submodule _run_on_acc_26

split_name=_run_on_acc_26, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_26, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_150 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_150 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_151 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_151 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_152 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_152 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_153 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_153 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.031840

TRT INetwork construction elapsed time: 0:00:00.031840

Build TRT engine elapsed time: 0:00:00.403447

Build TRT engine elapsed time: 0:00:00.403447

Lowering submodule _run_on_acc_26 elapsed time 0:00:00.465066

Lowering submodule _run_on_acc_26 elapsed time 0:00:00.465066

Now lowering submodule _run_on_acc_28

Now lowering submodule _run_on_acc_28

split_name=_run_on_acc_28, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_28, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_154 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_154 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_155 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_155 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.020004

TRT INetwork construction elapsed time: 0:00:00.020004

2023-03-27 05:21:03.309

Build TRT engine elapsed time: 0:00:04.404170

Build TRT engine elapsed time: 0:00:04.404170

Lowering submodule _run_on_acc_28 elapsed time 0:00:04.457413

Lowering submodule _run_on_acc_28 elapsed time 0:00:04.457413

Now lowering submodule _run_on_acc_30

Now lowering submodule _run_on_acc_30

split_name=_run_on_acc_30, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_30, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:21:03.352

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:21:03.370

TRT INetwork construction elapsed time: 0:00:00.019124

TRT INetwork construction elapsed time: 0:00:00.019124

2023-03-27 05:21:13.399

Build TRT engine elapsed time: 0:00:10.022734

Build TRT engine elapsed time: 0:00:10.022734

Lowering submodule _run_on_acc_30 elapsed time 0:00:10.072859

Lowering submodule _run_on_acc_30 elapsed time 0:00:10.072859

Now lowering submodule _run_on_acc_32

Now lowering submodule _run_on_acc_32

split_name=_run_on_acc_32, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_32, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_156 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_156 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_157 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_157 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_158 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_158 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_159 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_159 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.033510

TRT INetwork construction elapsed time: 0:00:00.033510

Build TRT engine elapsed time: 0:00:00.410851

Build TRT engine elapsed time: 0:00:00.410851

Lowering submodule _run_on_acc_32 elapsed time 0:00:00.474359

Lowering submodule _run_on_acc_32 elapsed time 0:00:00.474359

Now lowering submodule _run_on_acc_34

Now lowering submodule _run_on_acc_34

split_name=_run_on_acc_34, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_34, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_160 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_160 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_161 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_161 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.021186

TRT INetwork construction elapsed time: 0:00:00.021186

2023-03-27 05:21:18.137

Build TRT engine elapsed time: 0:00:04.182155

Build TRT engine elapsed time: 0:00:04.182155

Lowering submodule _run_on_acc_34 elapsed time 0:00:04.237904

Lowering submodule _run_on_acc_34 elapsed time 0:00:04.237904

Now lowering submodule _run_on_acc_36

Now lowering submodule _run_on_acc_36

split_name=_run_on_acc_36, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_36, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:21:18.181

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:21:18.199

TRT INetwork construction elapsed time: 0:00:00.018492

TRT INetwork construction elapsed time: 0:00:00.018492

2023-03-27 05:21:28.173

Build TRT engine elapsed time: 0:00:09.965197

Build TRT engine elapsed time: 0:00:09.965197

Lowering submodule _run_on_acc_36 elapsed time 0:00:10.016874

Lowering submodule _run_on_acc_36 elapsed time 0:00:10.016874

Now lowering submodule _run_on_acc_38

Now lowering submodule _run_on_acc_38

split_name=_run_on_acc_38, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_38, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_162 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_162 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_163 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_163 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_164 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_164 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_165 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_165 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.035506

TRT INetwork construction elapsed time: 0:00:00.035506

Build TRT engine elapsed time: 0:00:00.409342

Build TRT engine elapsed time: 0:00:00.409342

Lowering submodule _run_on_acc_38 elapsed time 0:00:00.477063

Lowering submodule _run_on_acc_38 elapsed time 0:00:00.477063

Now lowering submodule _run_on_acc_40

Now lowering submodule _run_on_acc_40

split_name=_run_on_acc_40, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_40, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_166 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_166 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_167 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_167 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.021001

TRT INetwork construction elapsed time: 0:00:00.021001

2023-03-27 05:21:32.882

Build TRT engine elapsed time: 0:00:04.148708

Build TRT engine elapsed time: 0:00:04.148708

Lowering submodule _run_on_acc_40 elapsed time 0:00:04.206795

Lowering submodule _run_on_acc_40 elapsed time 0:00:04.206795

Now lowering submodule _run_on_acc_42

Now lowering submodule _run_on_acc_42

split_name=_run_on_acc_42, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_42, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:21:32.927

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:21:32.946

TRT INetwork construction elapsed time: 0:00:00.019506

TRT INetwork construction elapsed time: 0:00:00.019506

2023-03-27 05:21:42.970

Build TRT engine elapsed time: 0:00:10.017006

Build TRT engine elapsed time: 0:00:10.017006

Lowering submodule _run_on_acc_42 elapsed time 0:00:10.069570

Lowering submodule _run_on_acc_42 elapsed time 0:00:10.069570

Now lowering submodule _run_on_acc_44

Now lowering submodule _run_on_acc_44

split_name=_run_on_acc_44, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_44, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_168 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_168 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_169 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_169 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_170 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_170 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_171 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_171 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.034506

TRT INetwork construction elapsed time: 0:00:00.034506

Build TRT engine elapsed time: 0:00:00.427618

Build TRT engine elapsed time: 0:00:00.427618

Lowering submodule _run_on_acc_44 elapsed time 0:00:00.496209

Lowering submodule _run_on_acc_44 elapsed time 0:00:00.496209

Now lowering submodule _run_on_acc_46

Now lowering submodule _run_on_acc_46

split_name=_run_on_acc_46, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_46, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_172 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_172 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_173 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_173 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.021894

TRT INetwork construction elapsed time: 0:00:00.021894

2023-03-27 05:21:47.799

Build TRT engine elapsed time: 0:00:04.246928

Build TRT engine elapsed time: 0:00:04.246928

Lowering submodule _run_on_acc_46 elapsed time 0:00:04.305896

Lowering submodule _run_on_acc_46 elapsed time 0:00:04.305896

Now lowering submodule _run_on_acc_48

Now lowering submodule _run_on_acc_48

split_name=_run_on_acc_48, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_48, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:21:47.845

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:21:47.865

TRT INetwork construction elapsed time: 0:00:00.020405

TRT INetwork construction elapsed time: 0:00:00.020405

2023-03-27 05:21:58.114

Build TRT engine elapsed time: 0:00:10.241350

Build TRT engine elapsed time: 0:00:10.241350

Lowering submodule _run_on_acc_48 elapsed time 0:00:10.294995

Lowering submodule _run_on_acc_48 elapsed time 0:00:10.294995

Now lowering submodule _run_on_acc_50

Now lowering submodule _run_on_acc_50

split_name=_run_on_acc_50, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_50, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_174 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_174 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_175 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_175 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_176 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_176 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_177 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_177 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.035008

TRT INetwork construction elapsed time: 0:00:00.035008

Build TRT engine elapsed time: 0:00:00.426671

Build TRT engine elapsed time: 0:00:00.426671

Lowering submodule _run_on_acc_50 elapsed time 0:00:00.496741

Lowering submodule _run_on_acc_50 elapsed time 0:00:00.496741

Now lowering submodule _run_on_acc_52

Now lowering submodule _run_on_acc_52

split_name=_run_on_acc_52, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_52, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_178 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_178 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_179 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_179 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.022005

TRT INetwork construction elapsed time: 0:00:00.022005

2023-03-27 05:22:02.946

Build TRT engine elapsed time: 0:00:04.249548

Build TRT engine elapsed time: 0:00:04.249548

Lowering submodule _run_on_acc_52 elapsed time 0:00:04.307900

Lowering submodule _run_on_acc_52 elapsed time 0:00:04.307900

Now lowering submodule _run_on_acc_54

Now lowering submodule _run_on_acc_54

split_name=_run_on_acc_54, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_54, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:22:02.991

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:22:03.002

TRT INetwork construction elapsed time: 0:00:00.011003

TRT INetwork construction elapsed time: 0:00:00.011003

2023-03-27 05:22:13.196

Build TRT engine elapsed time: 0:00:10.185831

Build TRT engine elapsed time: 0:00:10.185831

Lowering submodule _run_on_acc_54 elapsed time 0:00:10.230870

Lowering submodule _run_on_acc_54 elapsed time 0:00:10.230870

Now lowering submodule _run_on_acc_56

Now lowering submodule _run_on_acc_56

split_name=_run_on_acc_56, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_56, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:22:13.238

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:22:13.247

TRT INetwork construction elapsed time: 0:00:00.010002

TRT INetwork construction elapsed time: 0:00:00.010002

2023-03-27 05:22:16.968

Build TRT engine elapsed time: 0:00:03.713161

Build TRT engine elapsed time: 0:00:03.713161

Lowering submodule _run_on_acc_56 elapsed time 0:00:03.757133

Lowering submodule _run_on_acc_56 elapsed time 0:00:03.757133

Now lowering submodule _run_on_acc_58

Now lowering submodule _run_on_acc_58

split_name=_run_on_acc_58, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_58, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_180 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_180 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_181 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_181 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_182 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_182 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_183 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_183 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.036506

TRT INetwork construction elapsed time: 0:00:00.036506

Build TRT engine elapsed time: 0:00:00.424042

Build TRT engine elapsed time: 0:00:00.424042

Lowering submodule _run_on_acc_58 elapsed time 0:00:00.494137

Lowering submodule _run_on_acc_58 elapsed time 0:00:00.494137

Now lowering submodule _run_on_acc_60

Now lowering submodule _run_on_acc_60

split_name=_run_on_acc_60, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_60, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_184 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_184 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_185 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_185 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.021507

TRT INetwork construction elapsed time: 0:00:00.021507

2023-03-27 05:22:21.761

Build TRT engine elapsed time: 0:00:04.211298

Build TRT engine elapsed time: 0:00:04.211298

Lowering submodule _run_on_acc_60 elapsed time 0:00:04.271451

Lowering submodule _run_on_acc_60 elapsed time 0:00:04.271451

Now lowering submodule _run_on_acc_62

Now lowering submodule _run_on_acc_62

split_name=_run_on_acc_62, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_62, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:22:21.808

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:22:21.828

TRT INetwork construction elapsed time: 0:00:00.020001

TRT INetwork construction elapsed time: 0:00:00.020001

2023-03-27 05:22:31.858

Build TRT engine elapsed time: 0:00:10.021450

Build TRT engine elapsed time: 0:00:10.021450

Lowering submodule _run_on_acc_62 elapsed time 0:00:10.077401

Lowering submodule _run_on_acc_62 elapsed time 0:00:10.077401

Now lowering submodule _run_on_acc_64

Now lowering submodule _run_on_acc_64

split_name=_run_on_acc_64, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_64, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_186 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_186 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_187 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_187 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_188 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_188 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_189 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_189 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.036505

TRT INetwork construction elapsed time: 0:00:00.036505

Build TRT engine elapsed time: 0:00:00.421742

Build TRT engine elapsed time: 0:00:00.421742

Lowering submodule _run_on_acc_64 elapsed time 0:00:00.493351

Lowering submodule _run_on_acc_64 elapsed time 0:00:00.493351

Now lowering submodule _run_on_acc_66

Now lowering submodule _run_on_acc_66

split_name=_run_on_acc_66, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_66, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_190 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_190 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_191 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_191 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.021504

TRT INetwork construction elapsed time: 0:00:00.021504

2023-03-27 05:22:36.642

Build TRT engine elapsed time: 0:00:04.201128

Build TRT engine elapsed time: 0:00:04.201128

Lowering submodule _run_on_acc_66 elapsed time 0:00:04.260797

Lowering submodule _run_on_acc_66 elapsed time 0:00:04.260797

Now lowering submodule _run_on_acc_68

Now lowering submodule _run_on_acc_68

split_name=_run_on_acc_68, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_68, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:22:36.689

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:22:36.708

TRT INetwork construction elapsed time: 0:00:00.020005

TRT INetwork construction elapsed time: 0:00:00.020005

2023-03-27 05:22:46.710

Build TRT engine elapsed time: 0:00:09.994147

Build TRT engine elapsed time: 0:00:09.994147

Lowering submodule _run_on_acc_68 elapsed time 0:00:10.048280

Lowering submodule _run_on_acc_68 elapsed time 0:00:10.048280

Now lowering submodule _run_on_acc_70

Now lowering submodule _run_on_acc_70

split_name=_run_on_acc_70, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_70, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_192 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_192 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_193 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_193 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_194 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_194 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_195 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_195 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.037005

TRT INetwork construction elapsed time: 0:00:00.037005

Build TRT engine elapsed time: 0:00:00.435881

Build TRT engine elapsed time: 0:00:00.435881

Lowering submodule _run_on_acc_70 elapsed time 0:00:00.507967

Lowering submodule _run_on_acc_70 elapsed time 0:00:00.507967

Now lowering submodule _run_on_acc_72

Now lowering submodule _run_on_acc_72

split_name=_run_on_acc_72, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_72, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_196 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_196 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_197 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_197 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.022005

TRT INetwork construction elapsed time: 0:00:00.022005

2023-03-27 05:22:51.571

Build TRT engine elapsed time: 0:00:04.264076

Build TRT engine elapsed time: 0:00:04.264076

Lowering submodule _run_on_acc_72 elapsed time 0:00:04.324360

Lowering submodule _run_on_acc_72 elapsed time 0:00:04.324360

Now lowering submodule _run_on_acc_74

Now lowering submodule _run_on_acc_74

split_name=_run_on_acc_74, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_74, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:22:51.618

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:22:51.637

TRT INetwork construction elapsed time: 0:00:00.020490

TRT INetwork construction elapsed time: 0:00:00.020490

2023-03-27 05:23:01.792

Build TRT engine elapsed time: 0:00:10.146359

Build TRT engine elapsed time: 0:00:10.146359

Lowering submodule _run_on_acc_74 elapsed time 0:00:10.201391

Lowering submodule _run_on_acc_74 elapsed time 0:00:10.201391

Now lowering submodule _run_on_acc_76

Now lowering submodule _run_on_acc_76

split_name=_run_on_acc_76, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_76, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_198 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_198 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_199 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_199 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_200 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_200 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_201 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_201 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.037009

TRT INetwork construction elapsed time: 0:00:00.037009

Build TRT engine elapsed time: 0:00:00.419433

Build TRT engine elapsed time: 0:00:00.419433

Lowering submodule _run_on_acc_76 elapsed time 0:00:00.491929

Lowering submodule _run_on_acc_76 elapsed time 0:00:00.491929

Now lowering submodule _run_on_acc_78

Now lowering submodule _run_on_acc_78

split_name=_run_on_acc_78, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_78, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_202 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_202 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_203 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_203 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.022006

TRT INetwork construction elapsed time: 0:00:00.022006

2023-03-27 05:23:06.598

Build TRT engine elapsed time: 0:00:04.224634

Build TRT engine elapsed time: 0:00:04.224634

Lowering submodule _run_on_acc_78 elapsed time 0:00:04.284197

Lowering submodule _run_on_acc_78 elapsed time 0:00:04.284197

Now lowering submodule _run_on_acc_80

Now lowering submodule _run_on_acc_80

split_name=_run_on_acc_80, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_80, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:23:06.644

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:23:06.664

TRT INetwork construction elapsed time: 0:00:00.019939

TRT INetwork construction elapsed time: 0:00:00.019939

2023-03-27 05:23:16.660

Build TRT engine elapsed time: 0:00:09.987888

Build TRT engine elapsed time: 0:00:09.987888

Lowering submodule _run_on_acc_80 elapsed time 0:00:10.041882

Lowering submodule _run_on_acc_80 elapsed time 0:00:10.041882

Now lowering submodule _run_on_acc_82

Now lowering submodule _run_on_acc_82

split_name=_run_on_acc_82, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_82, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_204 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_204 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_205 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_205 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_206 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_206 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_207 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_207 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.036506

TRT INetwork construction elapsed time: 0:00:00.036506

Build TRT engine elapsed time: 0:00:00.433769

Build TRT engine elapsed time: 0:00:00.433769

Lowering submodule _run_on_acc_82 elapsed time 0:00:00.504658

Lowering submodule _run_on_acc_82 elapsed time 0:00:00.504658

Now lowering submodule _run_on_acc_84

Now lowering submodule _run_on_acc_84

split_name=_run_on_acc_84, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_84, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_208 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_208 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_209 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_209 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.022474

TRT INetwork construction elapsed time: 0:00:00.022474

2023-03-27 05:23:21.530

Build TRT engine elapsed time: 0:00:04.274198

Build TRT engine elapsed time: 0:00:04.274198

Lowering submodule _run_on_acc_84 elapsed time 0:00:04.336163

Lowering submodule _run_on_acc_84 elapsed time 0:00:04.336163

Now lowering submodule _run_on_acc_86

Now lowering submodule _run_on_acc_86

split_name=_run_on_acc_86, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_86, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:23:21.577

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:23:21.597

TRT INetwork construction elapsed time: 0:00:00.020005

TRT INetwork construction elapsed time: 0:00:00.020005

2023-03-27 05:23:31.919

Build TRT engine elapsed time: 0:00:10.314745

Build TRT engine elapsed time: 0:00:10.314745

Lowering submodule _run_on_acc_86 elapsed time 0:00:10.369284

Lowering submodule _run_on_acc_86 elapsed time 0:00:10.369284

Now lowering submodule _run_on_acc_88

Now lowering submodule _run_on_acc_88

split_name=_run_on_acc_88, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_88, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_210 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_210 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_211 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_211 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_212 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_212 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_213 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_213 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.038002

TRT INetwork construction elapsed time: 0:00:00.038002

Build TRT engine elapsed time: 0:00:00.421442

Build TRT engine elapsed time: 0:00:00.421442

Lowering submodule _run_on_acc_88 elapsed time 0:00:00.493725

Lowering submodule _run_on_acc_88 elapsed time 0:00:00.493725

Now lowering submodule _run_on_acc_90

Now lowering submodule _run_on_acc_90

split_name=_run_on_acc_90, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_90, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_214 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_214 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_215 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_215 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.022619

TRT INetwork construction elapsed time: 0:00:00.022619

2023-03-27 05:23:36.797

Build TRT engine elapsed time: 0:00:04.292359

Build TRT engine elapsed time: 0:00:04.292359

Lowering submodule _run_on_acc_90 elapsed time 0:00:04.353155

Lowering submodule _run_on_acc_90 elapsed time 0:00:04.353155

Now lowering submodule _run_on_acc_92

Now lowering submodule _run_on_acc_92

split_name=_run_on_acc_92, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_92, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:23:36.843

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:23:36.855

TRT INetwork construction elapsed time: 0:00:00.012449

TRT INetwork construction elapsed time: 0:00:00.012449

2023-03-27 05:23:47.226

Build TRT engine elapsed time: 0:00:10.362828

Build TRT engine elapsed time: 0:00:10.362828

Lowering submodule _run_on_acc_92 elapsed time 0:00:10.409648

Lowering submodule _run_on_acc_92 elapsed time 0:00:10.409648

Now lowering submodule _run_on_acc_94

Now lowering submodule _run_on_acc_94

split_name=_run_on_acc_94, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_94, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:23:47.270

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:23:47.280

TRT INetwork construction elapsed time: 0:00:00.012002

TRT INetwork construction elapsed time: 0:00:00.012002

2023-03-27 05:23:51.052

Build TRT engine elapsed time: 0:00:03.763718

Build TRT engine elapsed time: 0:00:03.763718

Lowering submodule _run_on_acc_94 elapsed time 0:00:03.808704

Lowering submodule _run_on_acc_94 elapsed time 0:00:03.808704

Now lowering submodule _run_on_acc_96

Now lowering submodule _run_on_acc_96

split_name=_run_on_acc_96, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_96, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_216 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_216 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_217 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_217 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_218 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_218 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_219 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_219 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.037003

TRT INetwork construction elapsed time: 0:00:00.037003

Build TRT engine elapsed time: 0:00:00.428686

Build TRT engine elapsed time: 0:00:00.428686

Lowering submodule _run_on_acc_96 elapsed time 0:00:00.500206

Lowering submodule _run_on_acc_96 elapsed time 0:00:00.500206

Now lowering submodule _run_on_acc_98

Now lowering submodule _run_on_acc_98

split_name=_run_on_acc_98, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_98, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_220 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_220 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_221 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_221 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.021504

TRT INetwork construction elapsed time: 0:00:00.021504

2023-03-27 05:23:55.944

Build TRT engine elapsed time: 0:00:04.304132

Build TRT engine elapsed time: 0:00:04.304132

Lowering submodule _run_on_acc_98 elapsed time 0:00:04.363832

Lowering submodule _run_on_acc_98 elapsed time 0:00:04.363832

Now lowering submodule _run_on_acc_100

Now lowering submodule _run_on_acc_100

split_name=_run_on_acc_100, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_100, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:23:55.990

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:23:56.010

TRT INetwork construction elapsed time: 0:00:00.019997

TRT INetwork construction elapsed time: 0:00:00.019997

2023-03-27 05:24:06.075

Build TRT engine elapsed time: 0:00:10.056246

Build TRT engine elapsed time: 0:00:10.056246

Lowering submodule _run_on_acc_100 elapsed time 0:00:10.110658

Lowering submodule _run_on_acc_100 elapsed time 0:00:10.110658

Now lowering submodule _run_on_acc_102

Now lowering submodule _run_on_acc_102

split_name=_run_on_acc_102, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_102, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_222 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_222 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_223 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_223 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_224 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_224 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_225 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_225 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.036510

TRT INetwork construction elapsed time: 0:00:00.036510

Build TRT engine elapsed time: 0:00:00.433247

Build TRT engine elapsed time: 0:00:00.433247

Lowering submodule _run_on_acc_102 elapsed time 0:00:00.505180

Lowering submodule _run_on_acc_102 elapsed time 0:00:00.505180

Now lowering submodule _run_on_acc_104

Now lowering submodule _run_on_acc_104

split_name=_run_on_acc_104, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_104, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_226 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_226 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_227 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_227 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.022015

TRT INetwork construction elapsed time: 0:00:00.022015

2023-03-27 05:24:10.928

Build TRT engine elapsed time: 0:00:04.258981

Build TRT engine elapsed time: 0:00:04.258981

Lowering submodule _run_on_acc_104 elapsed time 0:00:04.319536

Lowering submodule _run_on_acc_104 elapsed time 0:00:04.319536

Now lowering submodule _run_on_acc_106

Now lowering submodule _run_on_acc_106

split_name=_run_on_acc_106, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_106, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:24:10.974

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:24:10.994

TRT INetwork construction elapsed time: 0:00:00.020005

TRT INetwork construction elapsed time: 0:00:00.020005

2023-03-27 05:24:21.030

Build TRT engine elapsed time: 0:00:10.027874

Build TRT engine elapsed time: 0:00:10.027874

Lowering submodule _run_on_acc_106 elapsed time 0:00:10.082504

Lowering submodule _run_on_acc_106 elapsed time 0:00:10.082504

Now lowering submodule _run_on_acc_108

Now lowering submodule _run_on_acc_108

split_name=_run_on_acc_108, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_108, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_228 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_228 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_229 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_229 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_230 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_230 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_231 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_231 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.037008

TRT INetwork construction elapsed time: 0:00:00.037008

Build TRT engine elapsed time: 0:00:00.423912

Build TRT engine elapsed time: 0:00:00.423912

Lowering submodule _run_on_acc_108 elapsed time 0:00:00.495086

Lowering submodule _run_on_acc_108 elapsed time 0:00:00.495086

Now lowering submodule _run_on_acc_110

Now lowering submodule _run_on_acc_110

split_name=_run_on_acc_110, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_110, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_232 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_232 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_233 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_233 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.021300

TRT INetwork construction elapsed time: 0:00:00.021300

2023-03-27 05:24:25.911

Build TRT engine elapsed time: 0:00:04.294626

Build TRT engine elapsed time: 0:00:04.294626

Lowering submodule _run_on_acc_110 elapsed time 0:00:04.355888

Lowering submodule _run_on_acc_110 elapsed time 0:00:04.355888

Now lowering submodule _run_on_acc_112

Now lowering submodule _run_on_acc_112

split_name=_run_on_acc_112, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_112, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:24:25.958

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:24:25.978

TRT INetwork construction elapsed time: 0:00:00.021005

TRT INetwork construction elapsed time: 0:00:00.021005

2023-03-27 05:24:36.012

Build TRT engine elapsed time: 0:00:10.025766

Build TRT engine elapsed time: 0:00:10.025766

Lowering submodule _run_on_acc_112 elapsed time 0:00:10.080759

Lowering submodule _run_on_acc_112 elapsed time 0:00:10.080759

Now lowering submodule _run_on_acc_114

Now lowering submodule _run_on_acc_114

split_name=_run_on_acc_114, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_114, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_234 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_234 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_235 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_235 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_236 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_236 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_237 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_237 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.037007

TRT INetwork construction elapsed time: 0:00:00.037007

Build TRT engine elapsed time: 0:00:00.416450

Build TRT engine elapsed time: 0:00:00.416450

Lowering submodule _run_on_acc_114 elapsed time 0:00:00.487694

Lowering submodule _run_on_acc_114 elapsed time 0:00:00.487694

Now lowering submodule _run_on_acc_116

Now lowering submodule _run_on_acc_116

split_name=_run_on_acc_116, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_116, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_238 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_238 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_239 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_239 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.022006

TRT INetwork construction elapsed time: 0:00:00.022006

2023-03-27 05:24:40.836

Build TRT engine elapsed time: 0:00:04.247610

Build TRT engine elapsed time: 0:00:04.247610

Lowering submodule _run_on_acc_116 elapsed time 0:00:04.308578

Lowering submodule _run_on_acc_116 elapsed time 0:00:04.308578

Now lowering submodule _run_on_acc_118

Now lowering submodule _run_on_acc_118

split_name=_run_on_acc_118, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_118, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:24:40.884

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:24:40.904

TRT INetwork construction elapsed time: 0:00:00.020048

TRT INetwork construction elapsed time: 0:00:00.020048

2023-03-27 05:24:50.877

Build TRT engine elapsed time: 0:00:09.964763

Build TRT engine elapsed time: 0:00:09.964763

Lowering submodule _run_on_acc_118 elapsed time 0:00:10.021839

Lowering submodule _run_on_acc_118 elapsed time 0:00:10.021839

Now lowering submodule _run_on_acc_120

Now lowering submodule _run_on_acc_120

split_name=_run_on_acc_120, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_120, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_240 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_240 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_241 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_241 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_242 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_242 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_243 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_243 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.037499

TRT INetwork construction elapsed time: 0:00:00.037499

Build TRT engine elapsed time: 0:00:00.431422

Build TRT engine elapsed time: 0:00:00.431422

Lowering submodule _run_on_acc_120 elapsed time 0:00:00.503167

Lowering submodule _run_on_acc_120 elapsed time 0:00:00.503167

Now lowering submodule _run_on_acc_122

Now lowering submodule _run_on_acc_122

split_name=_run_on_acc_122, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_122, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_244 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_244 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_245 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_245 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.022005

TRT INetwork construction elapsed time: 0:00:00.022005

2023-03-27 05:24:55.738

Build TRT engine elapsed time: 0:00:04.267363

Build TRT engine elapsed time: 0:00:04.267363

Lowering submodule _run_on_acc_122 elapsed time 0:00:04.327424

Lowering submodule _run_on_acc_122 elapsed time 0:00:04.327424

Now lowering submodule _run_on_acc_124

Now lowering submodule _run_on_acc_124

split_name=_run_on_acc_124, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_124, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:24:55.784

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:24:55.804

TRT INetwork construction elapsed time: 0:00:00.020000

TRT INetwork construction elapsed time: 0:00:00.020000

2023-03-27 05:25:05.842

Build TRT engine elapsed time: 0:00:10.029499

Build TRT engine elapsed time: 0:00:10.029499

Lowering submodule _run_on_acc_124 elapsed time 0:00:10.084028

Lowering submodule _run_on_acc_124 elapsed time 0:00:10.084028

Now lowering submodule _run_on_acc_126

Now lowering submodule _run_on_acc_126

split_name=_run_on_acc_126, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_126, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_246 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_246 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_247 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_247 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_248 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_248 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_249 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_249 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.037984

TRT INetwork construction elapsed time: 0:00:00.037984

Build TRT engine elapsed time: 0:00:00.428732

Build TRT engine elapsed time: 0:00:00.428732

Lowering submodule _run_on_acc_126 elapsed time 0:00:00.502204

Lowering submodule _run_on_acc_126 elapsed time 0:00:00.502204

Now lowering submodule _run_on_acc_128

Now lowering submodule _run_on_acc_128

split_name=_run_on_acc_128, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_128, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_250 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_250 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_251 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_251 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.022002

TRT INetwork construction elapsed time: 0:00:00.022002

2023-03-27 05:25:10.696

Build TRT engine elapsed time: 0:00:04.261908

Build TRT engine elapsed time: 0:00:04.261908

Lowering submodule _run_on_acc_128 elapsed time 0:00:04.321579

Lowering submodule _run_on_acc_128 elapsed time 0:00:04.321579

Now lowering submodule _run_on_acc_130

Now lowering submodule _run_on_acc_130

split_name=_run_on_acc_130, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_130, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:25:10.742

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:25:10.753

TRT INetwork construction elapsed time: 0:00:00.012003

TRT INetwork construction elapsed time: 0:00:00.012003

2023-03-27 05:25:20.927

Build TRT engine elapsed time: 0:00:10.165734

Build TRT engine elapsed time: 0:00:10.165734

Lowering submodule _run_on_acc_130 elapsed time 0:00:10.211746

Lowering submodule _run_on_acc_130 elapsed time 0:00:10.211746

Now lowering submodule _run_on_acc_132

Now lowering submodule _run_on_acc_132

split_name=_run_on_acc_132, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_132, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:25:20.972

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:25:20.984

TRT INetwork construction elapsed time: 0:00:00.013004

TRT INetwork construction elapsed time: 0:00:00.013004

2023-03-27 05:25:24.712

Build TRT engine elapsed time: 0:00:03.719135

Build TRT engine elapsed time: 0:00:03.719135

Lowering submodule _run_on_acc_132 elapsed time 0:00:03.767137

Lowering submodule _run_on_acc_132 elapsed time 0:00:03.767137

Now lowering submodule _run_on_acc_134

Now lowering submodule _run_on_acc_134

split_name=_run_on_acc_134, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_134, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_252 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_252 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_253 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_253 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_254 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_254 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_255 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_255 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.038008

TRT INetwork construction elapsed time: 0:00:00.038008

Build TRT engine elapsed time: 0:00:00.443596

Build TRT engine elapsed time: 0:00:00.443596

Lowering submodule _run_on_acc_134 elapsed time 0:00:00.517405

Lowering submodule _run_on_acc_134 elapsed time 0:00:00.517405

Now lowering submodule _run_on_acc_136

Now lowering submodule _run_on_acc_136

split_name=_run_on_acc_136, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_136, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_256 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_256 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_257 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_257 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.022005

TRT INetwork construction elapsed time: 0:00:00.022005

2023-03-27 05:25:29.560

Build TRT engine elapsed time: 0:00:04.241779

Build TRT engine elapsed time: 0:00:04.241779

Lowering submodule _run_on_acc_136 elapsed time 0:00:04.302936

Lowering submodule _run_on_acc_136 elapsed time 0:00:04.302936

Now lowering submodule _run_on_acc_138

Now lowering submodule _run_on_acc_138

split_name=_run_on_acc_138, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_138, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:25:29.607

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:25:29.627

TRT INetwork construction elapsed time: 0:00:00.021001

TRT INetwork construction elapsed time: 0:00:00.021001

2023-03-27 05:25:39.760

Build TRT engine elapsed time: 0:00:10.125096

Build TRT engine elapsed time: 0:00:10.125096

Lowering submodule _run_on_acc_138 elapsed time 0:00:10.180046

Lowering submodule _run_on_acc_138 elapsed time 0:00:10.180046

Now lowering submodule _run_on_acc_140

Now lowering submodule _run_on_acc_140

split_name=_run_on_acc_140, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_140, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_258 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_258 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_259 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_259 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_260 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_260 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_261 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_261 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.036506

TRT INetwork construction elapsed time: 0:00:00.036506

Build TRT engine elapsed time: 0:00:00.428635

Build TRT engine elapsed time: 0:00:00.428635

Lowering submodule _run_on_acc_140 elapsed time 0:00:00.501459

Lowering submodule _run_on_acc_140 elapsed time 0:00:00.501459

Now lowering submodule _run_on_acc_142

Now lowering submodule _run_on_acc_142

split_name=_run_on_acc_142, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_142, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_262 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_262 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_263 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_263 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.022298

TRT INetwork construction elapsed time: 0:00:00.022298

2023-03-27 05:25:44.719

Build TRT engine elapsed time: 0:00:04.366112

Build TRT engine elapsed time: 0:00:04.366112

Lowering submodule _run_on_acc_142 elapsed time 0:00:04.427833

Lowering submodule _run_on_acc_142 elapsed time 0:00:04.427833

Now lowering submodule _run_on_acc_144

Now lowering submodule _run_on_acc_144

split_name=_run_on_acc_144, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_144, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:25:44.767

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:25:44.787

TRT INetwork construction elapsed time: 0:00:00.020982

TRT INetwork construction elapsed time: 0:00:00.020982

2023-03-27 05:25:54.843

Build TRT engine elapsed time: 0:00:10.047955

Build TRT engine elapsed time: 0:00:10.047955

Lowering submodule _run_on_acc_144 elapsed time 0:00:10.106318

Lowering submodule _run_on_acc_144 elapsed time 0:00:10.106318

Now lowering submodule _run_on_acc_146

Now lowering submodule _run_on_acc_146

split_name=_run_on_acc_146, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_146, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_264 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_264 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_265 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_265 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_266 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_266 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_267 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_267 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.037005

TRT INetwork construction elapsed time: 0:00:00.037005

Build TRT engine elapsed time: 0:00:00.431744

Build TRT engine elapsed time: 0:00:00.431744

Lowering submodule _run_on_acc_146 elapsed time 0:00:00.503690

Lowering submodule _run_on_acc_146 elapsed time 0:00:00.503690

Now lowering submodule _run_on_acc_148

Now lowering submodule _run_on_acc_148

split_name=_run_on_acc_148, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_148, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_268 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_268 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_269 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_269 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.022503

TRT INetwork construction elapsed time: 0:00:00.022503

2023-03-27 05:25:59.737

Build TRT engine elapsed time: 0:00:04.296204

Build TRT engine elapsed time: 0:00:04.296204

Lowering submodule _run_on_acc_148 elapsed time 0:00:04.356075

Lowering submodule _run_on_acc_148 elapsed time 0:00:04.356075

Now lowering submodule _run_on_acc_150

Now lowering submodule _run_on_acc_150

split_name=_run_on_acc_150, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_150, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:25:59.783

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:25:59.803

TRT INetwork construction elapsed time: 0:00:00.021011

TRT INetwork construction elapsed time: 0:00:00.021011

2023-03-27 05:26:09.847

Build TRT engine elapsed time: 0:00:10.035541

Build TRT engine elapsed time: 0:00:10.035541

Lowering submodule _run_on_acc_150 elapsed time 0:00:10.092231

Lowering submodule _run_on_acc_150 elapsed time 0:00:10.092231

Now lowering submodule _run_on_acc_152

Now lowering submodule _run_on_acc_152

split_name=_run_on_acc_152, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_152, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_270 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_270 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_271 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_271 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_272 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_272 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_273 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_273 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.038792

TRT INetwork construction elapsed time: 0:00:00.038792

Build TRT engine elapsed time: 0:00:00.436945

Build TRT engine elapsed time: 0:00:00.436945

Lowering submodule _run_on_acc_152 elapsed time 0:00:00.511085

Lowering submodule _run_on_acc_152 elapsed time 0:00:00.511085

Now lowering submodule _run_on_acc_154

Now lowering submodule _run_on_acc_154

split_name=_run_on_acc_154, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_154, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_274 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_274 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_275 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_275 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.023795

TRT INetwork construction elapsed time: 0:00:00.023795

2023-03-27 05:26:14.778

Build TRT engine elapsed time: 0:00:04.324144

Build TRT engine elapsed time: 0:00:04.324144

Lowering submodule _run_on_acc_154 elapsed time 0:00:04.387566

Lowering submodule _run_on_acc_154 elapsed time 0:00:04.387566

Now lowering submodule _run_on_acc_156

Now lowering submodule _run_on_acc_156

split_name=_run_on_acc_156, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_156, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:26:14.825

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:26:14.845

TRT INetwork construction elapsed time: 0:00:00.020507

TRT INetwork construction elapsed time: 0:00:00.020507

2023-03-27 05:26:24.844

Build TRT engine elapsed time: 0:00:09.991012

Build TRT engine elapsed time: 0:00:09.991012

Lowering submodule _run_on_acc_156 elapsed time 0:00:10.049578

Lowering submodule _run_on_acc_156 elapsed time 0:00:10.049578

Now lowering submodule _run_on_acc_158

Now lowering submodule _run_on_acc_158

split_name=_run_on_acc_158, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_158, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_276 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_276 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_277 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_277 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_278 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_278 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_279 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_279 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.037005

TRT INetwork construction elapsed time: 0:00:00.037005

Build TRT engine elapsed time: 0:00:00.423171

Build TRT engine elapsed time: 0:00:00.423171

Lowering submodule _run_on_acc_158 elapsed time 0:00:00.495677

Lowering submodule _run_on_acc_158 elapsed time 0:00:00.495677

Now lowering submodule _run_on_acc_160

Now lowering submodule _run_on_acc_160

split_name=_run_on_acc_160, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_160, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_280 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_280 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_281 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_281 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.021485

TRT INetwork construction elapsed time: 0:00:00.021485

2023-03-27 05:26:29.713

Build TRT engine elapsed time: 0:00:04.278824

Build TRT engine elapsed time: 0:00:04.278824

Lowering submodule _run_on_acc_160 elapsed time 0:00:04.340067

Lowering submodule _run_on_acc_160 elapsed time 0:00:04.340067

Now lowering submodule _run_on_acc_162

Now lowering submodule _run_on_acc_162

split_name=_run_on_acc_162, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_162, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:26:29.760

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:26:29.780

TRT INetwork construction elapsed time: 0:00:00.020991

TRT INetwork construction elapsed time: 0:00:00.020991

2023-03-27 05:26:39.806

Build TRT engine elapsed time: 0:00:10.018291

Build TRT engine elapsed time: 0:00:10.018291

Lowering submodule _run_on_acc_162 elapsed time 0:00:10.073605

Lowering submodule _run_on_acc_162 elapsed time 0:00:10.073605

Now lowering submodule _run_on_acc_164

Now lowering submodule _run_on_acc_164

split_name=_run_on_acc_164, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_164, input_specs=[InputTensorSpec(shape=torch.Size([1, 48, 80, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 48, 80, 1]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_282 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_282 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_283 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_283 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_284 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_284 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_285 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_285 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.038008

TRT INetwork construction elapsed time: 0:00:00.038008

Build TRT engine elapsed time: 0:00:00.424398

Build TRT engine elapsed time: 0:00:00.424398

Lowering submodule _run_on_acc_164 elapsed time 0:00:00.496390

Lowering submodule _run_on_acc_164 elapsed time 0:00:00.496390

Now lowering submodule _run_on_acc_166

Now lowering submodule _run_on_acc_166

split_name=_run_on_acc_166, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_166, input_specs=[InputTensorSpec(shape=torch.Size([60, 64, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([60, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 60, 1, 64, 64]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_286 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_286 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_287 are constant. In this case, please consider constant fold the model first.
warnings.warn(

I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\converters\converter_utils.py:457: UserWarning: Both operands of the binary elementwise op floordiv_287 are constant. In this case, please consider constant fold the model first.
warnings.warn(

TRT INetwork construction elapsed time: 0:00:00.023005

TRT INetwork construction elapsed time: 0:00:00.023005

2023-03-27 05:26:44.701

Build TRT engine elapsed time: 0:00:04.307120

Build TRT engine elapsed time: 0:00:04.307120

Lowering submodule _run_on_acc_166 elapsed time 0:00:04.368131

Lowering submodule _run_on_acc_166 elapsed time 0:00:04.368131

Now lowering submodule _run_on_acc_168

Now lowering submodule _run_on_acc_168

split_name=_run_on_acc_168, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_168, input_specs=[InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

2023-03-27 05:26:44.749

Unable to find layer norm plugin, fall back to TensorRT implementation.

Unable to find layer norm plugin, fall back to TensorRT implementation.

2023-03-27 05:26:44.760

TRT INetwork construction elapsed time: 0:00:00.012003

TRT INetwork construction elapsed time: 0:00:00.012003

2023-03-27 05:26:54.941

Build TRT engine elapsed time: 0:00:10.173015

Build TRT engine elapsed time: 0:00:10.173015

Lowering submodule _run_on_acc_168 elapsed time 0:00:10.220081

Lowering submodule _run_on_acc_168 elapsed time 0:00:10.220081

Now lowering submodule _run_on_acc_170

Now lowering submodule _run_on_acc_170

split_name=_run_on_acc_170, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_170, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 3840, 256]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.000999

TRT INetwork construction elapsed time: 0:00:00.000999

2023-03-27 05:26:57.108

Build TRT engine elapsed time: 0:00:02.113329

Build TRT engine elapsed time: 0:00:02.113329

Lowering submodule _run_on_acc_170 elapsed time 0:00:02.149360

Lowering submodule _run_on_acc_170 elapsed time 0:00:02.149360

Now lowering submodule _run_on_acc_172

Now lowering submodule _run_on_acc_172

split_name=_run_on_acc_172, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_172, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.002000

TRT INetwork construction elapsed time: 0:00:00.002000

Build TRT engine elapsed time: 0:00:01.918650

Build TRT engine elapsed time: 0:00:01.918650

Lowering submodule _run_on_acc_172 elapsed time 0:00:01.954914

Lowering submodule _run_on_acc_172 elapsed time 0:00:01.954914

Now lowering submodule _run_on_acc_174

Now lowering submodule _run_on_acc_174

split_name=_run_on_acc_174, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 512, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_174, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 512, 48, 80]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.003001

TRT INetwork construction elapsed time: 0:00:00.003001

2023-03-27 05:27:01.856

Build TRT engine elapsed time: 0:00:02.722756

Build TRT engine elapsed time: 0:00:02.722756

Lowering submodule _run_on_acc_174 elapsed time 0:00:02.760268

Lowering submodule _run_on_acc_174 elapsed time 0:00:02.760268

Now lowering submodule _run_on_acc_176

Now lowering submodule _run_on_acc_176

split_name=_run_on_acc_176, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_176, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.004177

TRT INetwork construction elapsed time: 0:00:00.004177

2023-03-27 05:27:04.652

Build TRT engine elapsed time: 0:00:02.740045

Build TRT engine elapsed time: 0:00:02.740045

Lowering submodule _run_on_acc_176 elapsed time 0:00:02.779477

Lowering submodule _run_on_acc_176 elapsed time 0:00:02.779477

Now lowering submodule _run_on_acc_178

Now lowering submodule _run_on_acc_178

split_name=_run_on_acc_178, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_178, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.003001

TRT INetwork construction elapsed time: 0:00:00.003001

2023-03-27 05:27:07.478

Build TRT engine elapsed time: 0:00:02.772769

Build TRT engine elapsed time: 0:00:02.772769

Lowering submodule _run_on_acc_178 elapsed time 0:00:02.809798

Lowering submodule _run_on_acc_178 elapsed time 0:00:02.809798

Now lowering submodule _run_on_acc_180

Now lowering submodule _run_on_acc_180

split_name=_run_on_acc_180, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_180, input_specs=[InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.003998

TRT INetwork construction elapsed time: 0:00:00.003998

2023-03-27 05:27:12.314

Build TRT engine elapsed time: 0:00:04.781494

Build TRT engine elapsed time: 0:00:04.781494

Lowering submodule _run_on_acc_180 elapsed time 0:00:04.821787

Lowering submodule _run_on_acc_180 elapsed time 0:00:04.821787

Now lowering submodule _run_on_acc_182

Now lowering submodule _run_on_acc_182

split_name=_run_on_acc_182, input_specs=[InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_182, input_specs=[InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 256, 96, 160]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.004128

TRT INetwork construction elapsed time: 0:00:00.004128

2023-03-27 05:27:40.716

Build TRT engine elapsed time: 0:00:28.343696

Build TRT engine elapsed time: 0:00:28.343696

Lowering submodule _run_on_acc_182 elapsed time 0:00:28.385848

Lowering submodule _run_on_acc_182 elapsed time 0:00:28.385848

Now lowering submodule _run_on_acc_184

Now lowering submodule _run_on_acc_184

split_name=_run_on_acc_184, input_specs=[InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_184, input_specs=[InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.004128

TRT INetwork construction elapsed time: 0:00:00.004128

2023-03-27 05:27:43.880

Build TRT engine elapsed time: 0:00:03.104550

Build TRT engine elapsed time: 0:00:03.104550

Lowering submodule _run_on_acc_184 elapsed time 0:00:03.146690

Lowering submodule _run_on_acc_184 elapsed time 0:00:03.146690

Now lowering submodule _run_on_acc_186

Now lowering submodule _run_on_acc_186

split_name=_run_on_acc_186, input_specs=[InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_186, input_specs=[InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.002000

TRT INetwork construction elapsed time: 0:00:00.002000

2023-03-27 05:27:46.898

Build TRT engine elapsed time: 0:00:02.965473

Build TRT engine elapsed time: 0:00:02.965473

Lowering submodule _run_on_acc_186 elapsed time 0:00:03.002562

Lowering submodule _run_on_acc_186 elapsed time 0:00:03.002562

Now lowering submodule _run_on_acc_188

Now lowering submodule _run_on_acc_188

split_name=_run_on_acc_188, input_specs=[InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_188, input_specs=[InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.001505

TRT INetwork construction elapsed time: 0:00:00.001505

2023-03-27 05:27:49.880

Build TRT engine elapsed time: 0:00:02.930511

Build TRT engine elapsed time: 0:00:02.930511

Lowering submodule _run_on_acc_188 elapsed time 0:00:02.967487

Lowering submodule _run_on_acc_188 elapsed time 0:00:02.967487

Now lowering submodule _run_on_acc_190

Now lowering submodule _run_on_acc_190

split_name=_run_on_acc_190, input_specs=[InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_190, input_specs=[InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 128, 192, 320]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.002011

TRT INetwork construction elapsed time: 0:00:00.002011

2023-03-27 05:27:54.905

Build TRT engine elapsed time: 0:00:04.971792

Build TRT engine elapsed time: 0:00:04.971792

Lowering submodule _run_on_acc_190 elapsed time 0:00:05.010673

Lowering submodule _run_on_acc_190 elapsed time 0:00:05.010673

Now lowering submodule _run_on_acc_192

Now lowering submodule _run_on_acc_192

split_name=_run_on_acc_192, input_specs=[InputTensorSpec(shape=torch.Size([1, 64, 384, 640]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_192, input_specs=[InputTensorSpec(shape=torch.Size([1, 64, 384, 640]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.000998

TRT INetwork construction elapsed time: 0:00:00.000998

2023-03-27 05:27:57.915

Build TRT engine elapsed time: 0:00:02.957856

Build TRT engine elapsed time: 0:00:02.957856

Lowering submodule _run_on_acc_192 elapsed time 0:00:02.992994

Lowering submodule _run_on_acc_192 elapsed time 0:00:02.992994

Now lowering submodule _run_on_acc_194

Now lowering submodule _run_on_acc_194

split_name=_run_on_acc_194, input_specs=[InputTensorSpec(shape=torch.Size([1, 64, 384, 640]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 64, 384, 640]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_194, input_specs=[InputTensorSpec(shape=torch.Size([1, 64, 384, 640]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True), InputTensorSpec(shape=torch.Size([1, 64, 384, 640]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.000999

TRT INetwork construction elapsed time: 0:00:00.000999

2023-03-27 05:28:00.995

Build TRT engine elapsed time: 0:00:03.027250

Build TRT engine elapsed time: 0:00:03.027250

Lowering submodule _run_on_acc_194 elapsed time 0:00:03.063577

Lowering submodule _run_on_acc_194 elapsed time 0:00:03.063577

Now lowering submodule _run_on_acc_196

Now lowering submodule _run_on_acc_196

split_name=_run_on_acc_196, input_specs=[InputTensorSpec(shape=torch.Size([1, 64, 384, 640]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

split_name=_run_on_acc_196, input_specs=[InputTensorSpec(shape=torch.Size([1, 64, 384, 640]), dtype=torch.float16, device=device(type='cuda', index=0), shape_ranges=[], has_batch_dim=True)]

Timing cache is used!

Timing cache is used!

TRT INetwork construction elapsed time: 0:00:00.001001

TRT INetwork construction elapsed time: 0:00:00.001001

2023-03-27 05:28:01.054

Failed to evaluate the script:
Python exception:

Traceback (most recent call last):
File "src\cython\vapoursynth.pyx", line 2866, in vapoursynth._vpy_evaluate
File "src\cython\vapoursynth.pyx", line 2867, in vapoursynth._vpy_evaluate
File "J:\tmp\tempPreviewVapoursynthFile05_19_33_068.vpy", line 38, in
clip = FeMaSR(clip=clip, device_index=0, trt=True, trt_cache_path=r"J:\tmp") # 640x352
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch\autograd\grad_mode.py", line 27, in decorate_context
return func(*args, **kwargs)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\vsfemasr\__init__.py", line 171, in femasr
module = lowerer(
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\lower.py", line 323, in __call__
return do_lower(module, inputs)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\passes\pass_utils.py", line 117, in pass_with_validation
processed_module = pass_(module, input, *args, **kwargs)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\lower.py", line 320, in do_lower
lower_result = pm(module)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch\fx\passes\pass_manager.py", line 240, in __call__
out = _pass(out)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch\fx\passes\pass_manager.py", line 240, in __call__
out = _pass(out)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\passes\lower_pass_manager_builder.py", line 167, in lower_func
lowered_module = self._lower_func(
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\lower.py", line 180, in lower_pass
interp_res: TRTInterpreterResult = interpreter(mod, input, module_name)
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\lower.py", line 132, in __call__
interp_result: TRTInterpreterResult = interpreter.run(
File "I:\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch_tensorrt\fx\fx2trt.py", line 252, in run
assert engine
AssertionError

=> I'll try whether going back to an older driver version helps (after work today).

HolyWu commented 1 year ago

Your log seems weird to me because almost every sentence repeats twice. Also please use Pastebin or Gist to post such a long log so as to let people view conversation much easier.

By the way, I could successfully build TensorRT engine on a 320x240 clip with my RTX 3050 (8GB VRAM, Game Ready driver 531.41). The build log is https://pastebin.com/wr092PqX.

import vapoursynth as vs
from vsfemasr import femasr

core = vs.core

clip = core.std.BlankClip(width=320, height=240, format=vs.RGBH, length=1)
clip = femasr(clip, trt=True)
clip.set_output()

However, on a 640x360 clip it failed quite quickly. The build log is https://pastebin.com/AKEv2bWL.

Selur commented 1 year ago

even when using:

# Imports
import vapoursynth as vs
# getting Vapoursynth core
core = vs.core
import site
import os
import ctypes
# Adding torch dependencies to PATH
path = site.getsitepackages()[0]+'/torch_dependencies/bin/'
ctypes.windll.kernel32.SetDllDirectoryW(path)
path = path.replace('\\', '/')
os.environ["PATH"] = path + os.pathsep + os.environ["PATH"]
os.environ["CUDA_MODULE_LOADING"] = "LAZY"

from vsfemasr import femasr
clip = core.std.BlankClip(width=320, height=240, format=vs.RGBH, length=1)
#clip = femasr(clip, trt=True)
clip.set_output()

I get the same issue. (https://pastebin.com/S1k050NB) (removing the CUDA_MODULE_LOADING-line doesn't change anything either) (vs-dpir works fine with trt, only femasr is having problems.) It even happens when I use clip = core.std.BlankClip(width=128, height=128, format=vs.RGBH, length=1)

Selur commented 1 year ago

Seeing Unable to find layer norm plugin, fall back to TensorRT implementation could I be missing some library?