Open azuryl opened 2 years ago
but squeezenet no problem torch2trt.tests.torchvision.classification.squeezenet1_0 | float16 | [(1, 3, 224, 224)] | {'fp16_mode': True} | 4.88E-04 | 82.79 | 4.89E-09 | 434 | 3.4e+03 | 2.26 | 0.314 | NUM_TESTS: 1 NUM_SUCCESSFUL_CONVERSION: 1 NUM_FAILED_CONVERSION: 0 NUM_ABOVE_TOLERANCE: 0 NUM_pSNR_TOLERANCE: 1
Hi,buddy,did u solve that problem,I meet the same problem as you. Thanks a lot if u can help me @azuryl
test alexnet [12/27/2021-23:27:14] [TRT] [E] 2: [ltWrapper.cpp::setupHeuristic::327] Error Code 2: Internal Error (Assertion cublasStatus == CUBLAS_STATUS_SUCCESS failed. ) | torch2trt.tests.torchvision.classification.alexnet | float16 | [(1, 3, 224, 224)] | {'fp16_mode': True} | N/A | N/A | N/A | N/A | N/A | Traceback (most recent call last): File "/home/delight-gpu/project/torch2trt/torch2trt/test.py", line 157, in
max_error,psnr_db,mse, fps, fps_trt, ms, ms_trt = run(test)
File "/home/delight-gpu/project/torch2trt/torch2trt/test.py", line 52, in run
outputs_trt = module_trt(inputs_trt)
File "/home/delight-gpu/anaconda3/envs/py397trt8221/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1102, in _call_impl
return forward_call(input, **kwargs)
File "/home/delight-gpu/project/torch2trt/torch2trt/torch2trt.py", line 466, in forward
idx = self.engine.get_binding_index(output_name)
AttributeError: 'NoneType' object has no attribute 'get_binding_index'
NUM_TESTS: 1 NUM_SUCCESSFUL_CONVERSION: 0 NUM_FAILED_CONVERSION: 1 NUM_ABOVE_TOLERANCE: 0 NUM_pSNR_TOLERANCE: 0