Closed MysticDaedra closed 1 month ago
Starting up SD.Next dev with --upgrade, presumably this error is from trying to load the selected unet-only Flux checkpoint?
The full error:
15:45:26-117627 INFO Loading UNet: type=FLUX file="D:\Stable Diffusion Files\Models\UNET\creartHyperFluxDevBnbNf4_hyperDevBnbNf4Unet.safetensors" offload=balanced quant=nf4 dtype=torch.float16 15:45:28-057356 ERROR Loading FLUX: Failed to convert UNET: split_with_sizes expects split_sizes to sum exactly to 33030144 (input tensor's size at dimension 0), but got split_sizes=[3072, 3072, 3072, 12288]
Using VENV: D:\automatic\venv 15:43:21-788203 INFO Starting SD.Next 15:43:21-792714 INFO Logger: file="D:\automatic\sdnext.log" level=DEBUG size=65 mode=create 15:43:21-793715 INFO Python version=3.11.8 platform=Windows bin="D:\automatic\venv\Scripts\Python.exe" venv="D:\automatic\venv" 15:43:24-246505 INFO Version: app=sd.next updated=2024-09-02 hash=045b24c0 branch=dev url=https://github.com/vladmandic/automatic/tree/dev ui=dev 15:43:24-816954 INFO Updating main repository 15:43:29-888223 INFO Upgraded to version: bba17766 Mon Sep 2 18:08:32 2024 -0400 15:43:29-895732 INFO Platform: arch=AMD64 cpu=AMD64 Family 25 Model 33 Stepping 2, AuthenticAMD system=Windows release=Windows-10-10.0.22631-SP0 python=3.11.8 15:43:29-896734 DEBUG Setting environment tuning 15:43:29-898733 INFO HF cache folder: C:\Users\Joshua.cache\huggingface\hub 15:43:29-898733 DEBUG Torch allocator: "garbage_collection_threshold:0.20,max_split_size_mb:512,backend:cudaMallocAsync" 15:43:29-916070 DEBUG Torch overrides: cuda=False rocm=False ipex=False diml=False openvino=False 15:43:29-917068 DEBUG Torch allowed: cuda=True rocm=True ipex=True diml=True openvino=True 15:43:29-930095 INFO nVidia CUDA toolkit detected: nvidia-smi present 15:43:30-152845 INFO Verifying requirements 15:43:30-155844 INFO Verifying packages 15:43:30-156845 INFO Startup: standard 15:43:30-157846 INFO Verifying submodules 15:43:32-916799 DEBUG Submodule: extensions-builtin/sd-extension-chainner / main 15:43:33-576603 DEBUG Submodule: extensions-builtin/sd-extension-system-info / main 15:43:34-219901 DEBUG Submodule: extensions-builtin/sd-webui-agent-scheduler / main 15:43:34-909362 DEBUG Submodule: extensions-builtin/sdnext-modernui / dev 15:43:35-617831 DEBUG Submodule: extensions-builtin/stable-diffusion-webui-rembg / master 15:43:36-315608 DEBUG Submodule: modules/k-diffusion / master 15:43:37-033878 DEBUG Git detached head detected: folder="wiki" reattach=master 15:43:37-034881 DEBUG Submodule: wiki / master 15:43:37-999493 DEBUG Register paths 15:43:38-059819 DEBUG Installed packages: 203 15:43:38-059819 DEBUG Extensions all: ['Lora', 'sd-extension-chainner', 'sd-extension-system-info', 'sd-webui-agent-scheduler', 'stable-diffusion-webui-images-browser', 'stable-diffusion-webui-rembg'] 15:43:38-677345 DEBUG Submodule: extensions-builtin\sd-extension-chainner / main 15:43:39-371360 DEBUG Submodule: extensions-builtin\sd-extension-system-info / main 15:43:40-057819 DEBUG Submodule: extensions-builtin\sd-webui-agent-scheduler / main 15:43:40-652624 DEBUG Running extension installer: D:\automatic\extensions-builtin\sd-webui-agent-scheduler\install.py 15:43:41-011305 DEBUG Submodule: extensions-builtin\stable-diffusion-webui-images-browser / main 15:43:41-699882 DEBUG Running extension installer: D:\automatic\extensions-builtin\stable-diffusion-webui-images-browser\install.py 15:43:42-069387 DEBUG Submodule: extensions-builtin\stable-diffusion-webui-rembg / master 15:43:42-679956 DEBUG Running extension installer: D:\automatic\extensions-builtin\stable-diffusion-webui-rembg\install.py 15:43:43-004336 DEBUG Extensions all: ['adetailer', 'canvas-zoom', 'sd-webui-infinite-image-browsing', 'sd-webui-temporal', 'ultimate-upscale-for-automatic1111'] 15:43:43-046972 DEBUG Submodule: extensions\adetailer / main 15:43:44-454354 DEBUG Running extension installer: D:\automatic\extensions\adetailer\install.py 15:43:44-837530 DEBUG Submodule: extensions\canvas-zoom / main 15:43:45-472476 DEBUG Running extension installer: D:\automatic\extensions\canvas-zoom\install.py 15:43:54-254284 DEBUG Submodule: extensions\sd-webui-infinite-image-browsing / main 15:43:54-899765 DEBUG Running extension installer: D:\automatic\extensions\sd-webui-infinite-image-browsing\install.py 15:44:02-696313 DEBUG Submodule: extensions\sd-webui-temporal / master 15:44:03-435476 DEBUG Submodule: extensions\ultimate-upscale-for-automatic1111 / master 15:44:04-108460 INFO Extensions enabled: ['Lora', 'sd-extension-chainner', 'sd-extension-system-info', 'sd-webui-agent-scheduler', 'stable-diffusion-webui-images-browser', 'stable-diffusion-webui-rembg', 'adetailer', 'canvas-zoom', 'sd-webui-infinite-image-browsing', 'sd-webui-temporal', 'ultimate-upscale-for-automatic1111'] 15:44:04-109778 INFO Verifying requirements 15:44:04-111296 INFO Updating Wiki 15:44:04-143396 DEBUG Submodule: D:\automatic\wiki / master 15:44:04-763404 DEBUG Setup complete without errors: 1725317045 15:44:04-768404 DEBUG Extension preload: {'extensions-builtin': 0.0, 'extensions': 0.0} 15:44:04-770404 DEBUG Starting module: <module 'webui' from 'D:\automatic\webui.py'> 15:44:04-771404 INFO Command line args: ['--debug', '--upgrade', '--share'] share=True upgrade=True debug=True 15:44:04-771910 DEBUG Env flags: ['SD_MASK_DEBUG=true'] 15:44:15-001322 INFO Load packages: {'torch': '2.3.1+cu121', 'diffusers': '0.30.2', 'gradio': '3.43.2'} 15:44:16-416527 INFO VRAM: Detected=8.0 GB Optimization=medvram 15:44:16-419528 DEBUG Read: file="config.json" json=81 bytes=5088 time=0.000 15:44:16-421033 DEBUG Unknown settings: ['cross_attention_options', 'ad_max_models', 'civitai_link_key', 'multiple_tqdm', 'ad_same_seed_for_each_tap', 'mudd_states', 'civitai_folder_lyco', 'diffusers_aesthetics_score', 'image_browser_active_tabs', 'ad_extra_models_dir', 'canvas_zoom_mask_clear', 'canvas_zoom_draw_staight_lines', 'ad_script_names'] 15:44:16-423039 INFO Engine: backend=Backend.DIFFUSERS compute=cuda device=cuda attention="Scaled-Dot-Product" mode=no_grad 15:44:16-439382 INFO Device: device=NVIDIA GeForce RTX 3070 n=1 arch=sm_90 cap=(8, 6) cuda=12.1 cudnn=8907 driver= 15:44:16-446897 DEBUG Read: file="html\reference.json" json=52 bytes=29118 time=0.006 15:44:16-962337 TRACE Trace: MASK 15:44:17-226029 DEBUG ONNX: version=1.18.0 provider=CUDAExecutionProvider, available=['TensorrtExecutionProvider', 'CUDAExecutionProvider', 'CPUExecutionProvider']
https://discord.com/channels/1101998836328697867/1130536562422186044/1280298597417484288
Diffusers
Standard
Dev
Other
duplicate of #3402
vladmandic
commented
1 month ago vladmandic
commented
1 month ago
Issue Description
Starting up SD.Next dev with --upgrade, presumably this error is from trying to load the selected unet-only Flux checkpoint?
The full error:
Version Platform Description
Using VENV: D:\automatic\venv 15:43:21-788203 INFO Starting SD.Next 15:43:21-792714 INFO Logger: file="D:\automatic\sdnext.log" level=DEBUG size=65 mode=create 15:43:21-793715 INFO Python version=3.11.8 platform=Windows bin="D:\automatic\venv\Scripts\Python.exe" venv="D:\automatic\venv" 15:43:24-246505 INFO Version: app=sd.next updated=2024-09-02 hash=045b24c0 branch=dev url=https://github.com/vladmandic/automatic/tree/dev ui=dev 15:43:24-816954 INFO Updating main repository 15:43:29-888223 INFO Upgraded to version: bba17766 Mon Sep 2 18:08:32 2024 -0400 15:43:29-895732 INFO Platform: arch=AMD64 cpu=AMD64 Family 25 Model 33 Stepping 2, AuthenticAMD system=Windows release=Windows-10-10.0.22631-SP0 python=3.11.8 15:43:29-896734 DEBUG Setting environment tuning 15:43:29-898733 INFO HF cache folder: C:\Users\Joshua.cache\huggingface\hub 15:43:29-898733 DEBUG Torch allocator: "garbage_collection_threshold:0.20,max_split_size_mb:512,backend:cudaMallocAsync" 15:43:29-916070 DEBUG Torch overrides: cuda=False rocm=False ipex=False diml=False openvino=False 15:43:29-917068 DEBUG Torch allowed: cuda=True rocm=True ipex=True diml=True openvino=True 15:43:29-930095 INFO nVidia CUDA toolkit detected: nvidia-smi present 15:43:30-152845 INFO Verifying requirements 15:43:30-155844 INFO Verifying packages 15:43:30-156845 INFO Startup: standard 15:43:30-157846 INFO Verifying submodules 15:43:32-916799 DEBUG Submodule: extensions-builtin/sd-extension-chainner / main 15:43:33-576603 DEBUG Submodule: extensions-builtin/sd-extension-system-info / main 15:43:34-219901 DEBUG Submodule: extensions-builtin/sd-webui-agent-scheduler / main 15:43:34-909362 DEBUG Submodule: extensions-builtin/sdnext-modernui / dev 15:43:35-617831 DEBUG Submodule: extensions-builtin/stable-diffusion-webui-rembg / master 15:43:36-315608 DEBUG Submodule: modules/k-diffusion / master 15:43:37-033878 DEBUG Git detached head detected: folder="wiki" reattach=master 15:43:37-034881 DEBUG Submodule: wiki / master 15:43:37-999493 DEBUG Register paths 15:43:38-059819 DEBUG Installed packages: 203 15:43:38-059819 DEBUG Extensions all: ['Lora', 'sd-extension-chainner', 'sd-extension-system-info', 'sd-webui-agent-scheduler', 'stable-diffusion-webui-images-browser', 'stable-diffusion-webui-rembg'] 15:43:38-677345 DEBUG Submodule: extensions-builtin\sd-extension-chainner / main 15:43:39-371360 DEBUG Submodule: extensions-builtin\sd-extension-system-info / main 15:43:40-057819 DEBUG Submodule: extensions-builtin\sd-webui-agent-scheduler / main 15:43:40-652624 DEBUG Running extension installer: D:\automatic\extensions-builtin\sd-webui-agent-scheduler\install.py 15:43:41-011305 DEBUG Submodule: extensions-builtin\stable-diffusion-webui-images-browser / main 15:43:41-699882 DEBUG Running extension installer: D:\automatic\extensions-builtin\stable-diffusion-webui-images-browser\install.py 15:43:42-069387 DEBUG Submodule: extensions-builtin\stable-diffusion-webui-rembg / master 15:43:42-679956 DEBUG Running extension installer: D:\automatic\extensions-builtin\stable-diffusion-webui-rembg\install.py 15:43:43-004336 DEBUG Extensions all: ['adetailer', 'canvas-zoom', 'sd-webui-infinite-image-browsing', 'sd-webui-temporal', 'ultimate-upscale-for-automatic1111'] 15:43:43-046972 DEBUG Submodule: extensions\adetailer / main 15:43:44-454354 DEBUG Running extension installer: D:\automatic\extensions\adetailer\install.py 15:43:44-837530 DEBUG Submodule: extensions\canvas-zoom / main 15:43:45-472476 DEBUG Running extension installer: D:\automatic\extensions\canvas-zoom\install.py 15:43:54-254284 DEBUG Submodule: extensions\sd-webui-infinite-image-browsing / main 15:43:54-899765 DEBUG Running extension installer: D:\automatic\extensions\sd-webui-infinite-image-browsing\install.py 15:44:02-696313 DEBUG Submodule: extensions\sd-webui-temporal / master 15:44:03-435476 DEBUG Submodule: extensions\ultimate-upscale-for-automatic1111 / master 15:44:04-108460 INFO Extensions enabled: ['Lora', 'sd-extension-chainner', 'sd-extension-system-info', 'sd-webui-agent-scheduler', 'stable-diffusion-webui-images-browser', 'stable-diffusion-webui-rembg', 'adetailer', 'canvas-zoom', 'sd-webui-infinite-image-browsing', 'sd-webui-temporal', 'ultimate-upscale-for-automatic1111'] 15:44:04-109778 INFO Verifying requirements 15:44:04-111296 INFO Updating Wiki 15:44:04-143396 DEBUG Submodule: D:\automatic\wiki / master 15:44:04-763404 DEBUG Setup complete without errors: 1725317045 15:44:04-768404 DEBUG Extension preload: {'extensions-builtin': 0.0, 'extensions': 0.0} 15:44:04-770404 DEBUG Starting module: <module 'webui' from 'D:\automatic\webui.py'> 15:44:04-771404 INFO Command line args: ['--debug', '--upgrade', '--share'] share=True upgrade=True debug=True 15:44:04-771910 DEBUG Env flags: ['SD_MASK_DEBUG=true'] 15:44:15-001322 INFO Load packages: {'torch': '2.3.1+cu121', 'diffusers': '0.30.2', 'gradio': '3.43.2'} 15:44:16-416527 INFO VRAM: Detected=8.0 GB Optimization=medvram 15:44:16-419528 DEBUG Read: file="config.json" json=81 bytes=5088 time=0.000 15:44:16-421033 DEBUG Unknown settings: ['cross_attention_options', 'ad_max_models', 'civitai_link_key', 'multiple_tqdm', 'ad_same_seed_for_each_tap', 'mudd_states', 'civitai_folder_lyco', 'diffusers_aesthetics_score', 'image_browser_active_tabs', 'ad_extra_models_dir', 'canvas_zoom_mask_clear', 'canvas_zoom_draw_staight_lines', 'ad_script_names'] 15:44:16-423039 INFO Engine: backend=Backend.DIFFUSERS compute=cuda device=cuda attention="Scaled-Dot-Product" mode=no_grad 15:44:16-439382 INFO Device: device=NVIDIA GeForce RTX 3070 n=1 arch=sm_90 cap=(8, 6) cuda=12.1 cudnn=8907 driver= 15:44:16-446897 DEBUG Read: file="html\reference.json" json=52 bytes=29118 time=0.006 15:44:16-962337 TRACE Trace: MASK 15:44:17-226029 DEBUG ONNX: version=1.18.0 provider=CUDAExecutionProvider, available=['TensorrtExecutionProvider', 'CUDAExecutionProvider', 'CPUExecutionProvider']
Relevant log output
Backend
Diffusers
UI
Standard
Branch
Dev
Model
Other
Acknowledgements