HSqure
commented
1 year ago Hi! When i run 'python quant.py --quant_mode test --subset_len 1 --batch_size 1 --deploy ',i get this error: [VAIQ_NOTE]: =>Quantizable module is generated.(quantize_result/Model.py)
[VAIQ_NOTE]: =>Get module with quantization. 200 0%| | 0/1 [00:00<?, ?it/s]/opt/vitis_ai/conda/envs/vitis-ai-pytorch/lib/python3.7/site-packages/torch/functional.py:445: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at /opt/conda/conda-bld/pytorch_1639180518675/work/aten/src/ATen/native/TensorShape.cpp:2157.) return _VF.meshgrid(tensors, **kwargs) # type: ignore[attr-defined] 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 3.40it/s] Scanning 'data/downloaded/val.cache' images and labels... 0 found, 200 missing, 0 empty, 0 corrupted: 100%|████████████████████| 200/200 [00:00<?, ?it/s] Class Images Targets P R mAP F1 Computing mAP: 0%| | 0/13 [00:00<?, ?it/s] [VAIQ_WARN]: The tensor type of Model::input_0 is torch.uint8. Only support float32/double quantization.
[VAIQ_WARN]: The tensor type of Model::Model/Focus[model]/Focus[0]/9546 is torch.uint8. Only support float32/double quantization.
[VAIQ_WARN]: The tensor type of Model::Model/Focus[model]/Focus[0]/9556 is torch.uint8. Only support float32/double quantization.
[VAIQ_WARN]: The tensor type of Model::Model/Focus[model]/Focus[0]/9566 is torch.uint8. Only support float32/double quantization.
[VAIQ_WARN]: The tensor type of Model::Model/Focus[model]/Focus[0]/9576 is torch.uint8. Only support float32/double quantization.
[VAIQ_WARN]: The tensor type of Model::Model/Focus[model]/Focus[0]/input.1 is torch.uint8. Only support float32/double quantization. Computing mAP: 0%| | 0/13 [00:00<?, ?it/s] Traceback (most recent call last): File "quant.py", line 505, in file_path=file_path) File "quant.py", line 473, in quantization register_buffers=register_buffers) File "quant.py", line 123, in test inf_out, train_out = model_with_post_precess(imgs, model, data_cfg, register_buffers) # inference and training outputs File "quant.py", line 354, in model_with_post_precess for output in model(images): File "/opt/vitis_ai/conda/envs/vitis-ai-pytorch/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1102, in _call_impl return forward_call(*input, *kwargs) File "quantize_result/Model.py", line 186, in forward output_module_1 = self.module_6(output_module_1) File "/opt/vitis_ai/conda/envs/vitis-ai-pytorch/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1120, in _call_impl result = forward_call(input, **kwargs) File "/opt/vitis_ai/conda/envs/vitis-ai-pytorch/lib/python3.7/site-packages/pytorch_nndct/nn/modules/conv.py", line 115, in forward groups = self.groups) RuntimeError: expected scalar type Byte but found Float can you help me?thank you!
Hi @shoayi ! Sorry It's been a long time so I forget some details about this program. Here I use some parameter directlty taking from model file may cause some compatbility issues.
Hi! When i run 'python quant.py --quant_mode test --subset_len 1 --batch_size 1 --deploy ',i get this error: [VAIQ_NOTE]: =>Quantizable module is generated.(quantize_result/Model.py)
[VAIQ_NOTE]: =>Get module with quantization. 200 0%| | 0/1 [00:00<?, ?it/s]/opt/vitis_ai/conda/envs/vitis-ai-pytorch/lib/python3.7/site-packages/torch/functional.py:445: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at /opt/conda/conda-bld/pytorch_1639180518675/work/aten/src/ATen/native/TensorShape.cpp:2157.) return _VF.meshgrid(tensors, **kwargs) # type: ignore[attr-defined] 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 3.40it/s] Scanning 'data/downloaded/val.cache' images and labels... 0 found, 200 missing, 0 empty, 0 corrupted: 100%|████████████████████| 200/200 [00:00<?, ?it/s] Class Images Targets P R mAP F1 Computing mAP: 0%| | 0/13 [00:00<?, ?it/s] [VAIQ_WARN]: The tensor type of Model::input_0 is torch.uint8. Only support float32/double quantization.
[VAIQ_WARN]: The tensor type of Model::Model/Focus[model]/Focus[0]/9546 is torch.uint8. Only support float32/double quantization.
[VAIQ_WARN]: The tensor type of Model::Model/Focus[model]/Focus[0]/9556 is torch.uint8. Only support float32/double quantization.
[VAIQ_WARN]: The tensor type of Model::Model/Focus[model]/Focus[0]/9566 is torch.uint8. Only support float32/double quantization.
[VAIQ_WARN]: The tensor type of Model::Model/Focus[model]/Focus[0]/9576 is torch.uint8. Only support float32/double quantization.
[VAIQ_WARN]: The tensor type of Model::Model/Focus[model]/Focus[0]/input.1 is torch.uint8. Only support float32/double quantization. Computing mAP: 0%| | 0/13 [00:00<?, ?it/s] Traceback (most recent call last): File "quant.py", line 505, in
file_path=file_path)
File "quant.py", line 473, in quantization
register_buffers=register_buffers)
File "quant.py", line 123, in test
inf_out, train_out = model_with_post_precess(imgs, model, data_cfg, register_buffers) # inference and training outputs
File "quant.py", line 354, in model_with_post_precess
for output in model(images):
File "/opt/vitis_ai/conda/envs/vitis-ai-pytorch/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1102, in _call_impl
return forward_call(*input, *kwargs)
File "quantize_result/Model.py", line 186, in forward
output_module_1 = self.module_6(output_module_1)
File "/opt/vitis_ai/conda/envs/vitis-ai-pytorch/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1120, in _call_impl
result = forward_call(input, **kwargs)
File "/opt/vitis_ai/conda/envs/vitis-ai-pytorch/lib/python3.7/site-packages/pytorch_nndct/nn/modules/conv.py", line 115, in forward
groups = self.groups)
RuntimeError: expected scalar type Byte but found Float
can you help me?thank you!