PPQ yolov5-7.0 ：return torch.cat(concat_view, axis=axis) RuntimeError: Sizes of tensors must match except in dimension 1. Expected size 28 but got size 27 for tensor number 1 in the list.

python ProgramEntrance_1.py 正准备量化你的网络，检查下列设置: WORKING DIRECTORY : working TARGET PLATFORM : TRT_INT8 NETWORK INPUTSHAPE : [1, 3, 640, 640] CALIBRATION BATCHSIZE: 16 [Warning] 你的输入图像似乎包含动态的尺寸，因此 CALIBRATION BATCHSIZE 被强制设置为 1 123 File(s) Loaded. Loaded sample 0, shape: torch.Size([1, 3, 426, 640]) Loaded sample 1, shape: torch.Size([1, 3, 640, 360]) Loaded sample 2, shape: torch.Size([1, 3, 480, 640]) Loaded sample 3, shape: torch.Size([1, 3, 480, 640]) Loaded sample 4, shape: torch.Size([1, 3, 425, 640]) Batch Shape: torch.Size([1, 3, 426, 640]) [Warning] Compling Kernels... Please wait (It will take a few minutes). 网络正量化中，根据你的量化配置，这将需要一段时间: [12:16:08] PPQ Quantization Fusion Pass Running ... Finished. [12:16:08] PPQ Quantize Simplify Pass Running ... Finished. [12:16:08] PPQ Parameter Quantization Pass Running ... Finished. Calibration Progress(Phase 1): 0%| | 0/32 [00:00<?, ?it/s] Traceback (most recent call last): File "/home/yellen/run/proj/03-zz-PPQ/ppq/ppq/executor/torch.py", line 536, in __forward outputs = operation_forward_func(operation, inputs, self._executing_context) File "/home/yellen/run/proj/03-zz-PPQ/ppq/ppq/executor/op/torch/default.py", line 901, in Concat_forward return torch.cat(concat_view, axis=axis) RuntimeError: Sizes of tensors must match except in dimension 1. Expected size 28 but got size 27 for tensor number 1 in the list.

The above exception was the direct cause of the following exception:

Traceback (most recent call last): File "/home/yellen/run/proj/03-zz-PPQ/ppq/ProgramEntrance_1.py", line 98, in quantized = quantize_native_model( File "/home/yellen/run/proj/03-zz-PPQ/ppq/ppq/core/defs.py", line 54, in _wrapper return func(*args, kwargs) File "/home/yellen/run/proj/03-zz-PPQ/ppq/ppq/api/interface.py", line 530, in quantize_native_model quantizer.quantize( File "/home/yellen/run/proj/03-zz-PPQ/ppq/ppq/core/defs.py", line 54, in _wrapper return func(args, kwargs) File "/home/yellen/run/proj/03-zz-PPQ/ppq/ppq/quantization/quantizer/base.py", line 69, in quantize quant_pipeline.optimize( File "/home/yellen/run/proj/03-zz-PPQ/ppq/ppq/quantization/optim/base.py", line 81, in optimize optim_pass.optimize(graph=graph, kwargs) File "/home/yellen/run/proj/03-zz-PPQ/ppq/ppq/core/defs.py", line 54, in _wrapper return func(args, kwargs) File "/home/yellen/run/proj/03-zz-PPQ/ppq/ppq/quantization/optim/calibration.py", line 176, in optimize self.calibrate(desc='Calibration Progress(Phase 1)', dataloader=dataloader, File "/home/yellen/run/proj/03-zz-PPQ/ppq/ppq/quantization/optim/calibration.py", line 117, in calibrate executor.forward(inputs=data, hooks=hooks, File "/home/yellen/miniconda3/envs/ppq/lib/python3.9/site-packages/torch/autograd/grad_mode.py", line 27, in decoratecontext return func(*args, **kwargs) File "/home/yellen/run/proj/03-zz-PPQ/ppq/ppq/executor/torch.py", line 405, in forward return self.forward( File "/home/yellen/run/proj/03-zz-PPQ/ppq/ppq/executor/torch.py", line 563, in forward raise RuntimeError(f'Op Execution Error: {str(operation)}') from RuntimeError: Op Execution Error: Concat_168(Type: Concat, Num of Input: 2, Num of Output: 1)

OpenPPL / ppq

PPQ yolov5-7.0 ：return torch.cat(concat_view, axis=axis) RuntimeError: Sizes of tensors must match except in dimension 1. Expected size 28 but got size 27 for tensor number 1 in the list. #520