Pytorch error: please report a bug to PyTorch

Hi All,

I am using wsl2 on windows 10 and using PyTorch versions as below:

pytorch 2.1.0 py3.11_cuda12.1_cudnn8.9.2_0 pytorch
pytorch-cuda 12.1 ha16c6d3_5 pytorch

I get the following error:

Starting Training Loop...
torch.Size([128, 3, 128, 128])
Traceback (most recent call last):
  File "/mnt/c/Users/hjakkamk/Downloads/2-ADRL/Assignment/assignment-2/assignment2.py", line 280, in <module>
    errD_real.backward()
  File "/home/hjakkamk/miniconda3/lib/python3.11/site-packages/torch/_tensor.py", line 492, in backward  
    torch.autograd.backward(
  File "/home/hjakkamk/miniconda3/lib/python3.11/site-packages/torch/autograd/__init__.py", line 251, in 
backward
    Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass       
RuntimeError: handle_0 INTERNAL ASSERT FAILED at "/opt/conda/conda-bld/pytorch_1695392035891/work/c10/cuda/driver_api.cpp":15, please report a bug to PyTorch.

When I run the code:

print("Starting Training Loop...")
# For each epoch
for epoch in range(num_epochs):
    # For each batch in the dataloader
    for i, data in enumerate(dataloader, 0):

        ############################
        # (1) Update D network: maximize log(D(x)) + log(1 - D(G(z)))
        ###########################
        ## Train with all-real batch
        netD.zero_grad()
        # Format batch
        real_cpu = data[0].to(device)
        print(real_cpu.shape)
        b_size = real_cpu.size(0)
        # print(b_size)
        label = torch.full((b_size,), real_label, dtype=torch.float, device=device)
        # Forward pass real batch through D
        output = netD(real_cpu).view(-1)
        # print(output.shape)
        # Calculate loss on all-real batch
        errD_real = criterion(output[:batch_size], label)
        # Calculate gradients for D in backward pass
        errD_real.backward()
        D_x = output.mean().item()

The error is at line errD_real.backward().

Appreciate any help on this. Thanks!

Could you create an issue on GitHub, please?

Can you please provide the exact repo where I need to create this issue? Thanks.

1 Like