search

ROCm / AITemplate

AITemplate is a Python framework which renders neural network into high performance CUDA/HIP C++ code. Specialized for FP16 TensorCore (NVIDIA GPU) and MatrixCore (AMD GPU) inference.
Apache License 2.0
11 stars 7 forks source link

test.so build faled #71

Closed Boom-Hacker closed 1 year ago

Boom-Hacker commented 1 year ago

2023-09-17 19:02:48,710 INFO generated 207 function srcs 2023-09-17 19:02:50,015 INFO generated 8 library srcs 2023-09-17 19:02:50,015 INFO Using 16 CPU for building 2023-09-17 19:07:31,399 INFO compiled the final .so file elapsed time: 0:04:41.384353 [19:07:31] model_container.cpp:67: Device Runtime Version: 50422803; Driver Version: 50422803 [19:07:31] model_container.cpp:81: Hardware accelerator device properties: Device: ASCII string identifying device: AMD Radeon RX 6750 XT Major compute capability: 10 Minor compute capability: 3 AMD GCN Arch Value: 1030 PCI bus ID of the device: 3 PCI device ID of the device: 0 Memory limits: Constant memory available on device in bytes: 2147483647 Global memory available on device in bytes: 12868124672 Global memory bus width in bits: 192 Size of L2 cache in bytes: 3145728 Shared memory available per block in bytes: 65536 Maximum Shared Memory Per Multiprocessor in bytes: 65536 [19:07:31] model_container.cpp:85: Init AITemplate Runtime with 1 concurrency Traceback (most recent call last): File "/home/beforespace/AITnew/AITemplate/examples/05_stable_diffusion/scripts/compile.py", line 106, in compile_diffusers() File "/home/beforespace/AITnew/AITemplate/venv/lib/python3.10/site-packages/click/core.py", line 1157, in call return self.main(args, kwargs) File "/home/beforespace/AITnew/AITemplate/venv/lib/python3.10/site-packages/click/core.py", line 1078, in main rv = self.invoke(ctx) File "/home/beforespace/AITnew/AITemplate/venv/lib/python3.10/site-packages/click/core.py", line 1434, in invoke return ctx.invoke(self.callback, ctx.params) File "/home/beforespace/AITnew/AITemplate/venv/lib/python3.10/site-packages/click/core.py", line 783, in invoke return __callback(args, *kwargs) File "/home/beforespace/AITnew/AITemplate/examples/05_stable_diffusion/scripts/compile.py", line 95, in compile_diffusers compile_vae( File "/home/beforespace/AITnew/AITemplate/examples/05_stable_diffusion/src/compile_lib/compile_vae.py", line 128, in compile_vae params_ait = map_vae_params(ait_vae, pt_mod, batch_size, height width) File "/home/beforespace/AITnew/AITemplate/examples/05_stable_diffusion/src/compile_lib/compile_vae.py", line 48, in map_vae_params q_weight = pt_params[prefix + "query.weight"] KeyError: 'decoder.mid_block.attentions.0.query.weight'

Boom-Hacker commented 1 year ago

use special version of diffusers and transformers,in 05 stable diffusion 's readme.md