An error on Colab Full Nerf

[lqy93]

When I run the Colab script for Full Nerf version an error occurred:

RuntimeError                              Traceback (most recent call last)
<ipython-input-26-1d7c06d853cb> in <module>()
    119     loss = coarse_loss + fine_loss
    120     print(coarse_loss.shape,fine_loss.shape)
--> 121     loss.backward()
    122     optimizer.step()
    123     optimizer.zero_grad()

1 frames
/usr/local/lib/python3.7/dist-packages/torch/_tensor.py in backward(self, gradient, retain_graph, create_graph, inputs)
    253                 create_graph=create_graph,
    254                 inputs=inputs)
--> 255         torch.autograd.backward(self, gradient, retain_graph, create_graph, inputs=inputs)
    256 
    257     def register_hook(self, hook):

/usr/local/lib/python3.7/dist-packages/torch/autograd/__init__.py in backward(tensors, grad_tensors, retain_graph, create_graph, grad_variables, inputs)
    147     Variable._execution_engine.run_backward(
    148         tensors, grad_tensors_, retain_graph, create_graph, inputs,
--> 149         allow_unreachable=True, accumulate_grad=True)  # allow_unreachable flag
    150 
    151 

**RuntimeError: Function AddmmBackward returned an invalid gradient at index 2 - got [72, 128] but expected shape compatible with [66, 128]**

I didn't change any code. How to solve the problem? Or do I need to set any configuration?

[curryJ]

Do you address this issue？
I didn't change any code too,but get another issue

[rajatthakur]

Uncommenting ReplicateNeRFModel fixed this issue for me.

model_coarse = ReplicateNeRFModel(
    hidden_size=128,
    num_encoding_fn_xyz=num_encoding_fn_xyz,
    num_encoding_fn_dir=num_encoding_fn_dir,
    include_input_xyz=include_input_xyz,
    include_input_dir=include_input_dir
)
# model_coarse = VeryTinyNeRFModel()
model_coarse.to(device)

# Initialize a fine-resolution model, if specified.
model_fine = ReplicateNeRFModel(
    hidden_size=128,
    num_encoding_fn_xyz=num_encoding_fn_xyz,
    num_encoding_fn_dir=num_encoding_fn_dir,
    include_input_xyz=include_input_xyz,
    include_input_dir=include_input_dir
)
# model_fine = VeryTinyNeRFModel()
model_fine.to(device)

[jdiazram]

I have the same error that @curryJ, but the solution of @rajatthakur is run now. Best