tendar

Hi, external_source_test may has a bug. The test file is https://github.com/NVIDIA/DALI/blob/main/dali/pipeline/operator/builtin/external_source_test.cc#L181 I think this test may has two place to fix. The first place is L181: tl_gpu_.Copy(tensor_list) is called batch_size...

Related to **Resnet50/pytorch** **Describe the bug** source code link https://github.com/NVIDIA/DeepLearningExamples/blob/master/PyTorch/Classification/ConvNets/image_classification/dataloaders.py#L153 If here define the pipeline with param dali_cpu=Ture, then the output of images = self.res(images) is already a gpu DataNode....