megum1 / beagle Goto Github PK

I encountered a runtime error while working on a project that involves performing element-wise operations between three tensors: batch_data, weights, and bias. The error message I received is as follows:
RuntimeError: The size of tensor a (3) must match the size of tensor b (0) at non-singleton dimension 1

The issue arises from a dimension mismatch along dimension 1 when trying to perform element-wise operations. Here are the shapes of the tensors involved:
batch_data shape: torch.Size([30, 3, 224, 224])
weights shape: torch.Size([30, 9, 224, 224])
bias shape: torch.Size([30, 3, 224, 224])

I have resolved the issue by using the bias tensor as it is, without dividing it into three parts. This approach correctly adds the bias term to each channel of the batch_data without any channel-wise adjustments, and it resolved the dimension mismatch error. If you have any further questions or encounter related issues, please feel free to reopen or create a new issue. Thank you for your support!