#Common Failure

#import Error

Error I: Cannot find the extension library(_C.so)

Solution:

• Make sure that the horizon_plugin_pytorch version corresponds to the cuda version.
• In python3, find the execution path of horizon_plugin_pytorch and check for .so files in that directory. There may be multiple versions of horizon_plugin_pytorch at the same time, so you need to uninstall it and keep only the one you need.

Error II: RuntimeError: Cannot load custom ops. Please rebuild the horizon_plugin_pytorch

Solution: check if the local CUDA environment is OK, such as path, version, etc.

#Module doesn't support deepcopy

Some frameworks (e.g., PyTorch Lightning) provide wrappers which do not support deepcopy around the native torch module.

Solution: Implement the __deepcopy__ method for the model.

class Model(Module):
    ...
    def __deepcopy__(self, memo):
        new_model = Model()
        new_model.xxx = self.xxx
        return new_model

#RuntimeError: Only Tensors created explicitly by the user (graph leaves) support the deepcopy protocol at the moment

This error usually occurs when deepcopying a torch.nn.Module. It typically happens when certain intermediate results are stored in the model during the forward pass, and the presence of non-leaf tensors in the model causes the deepcopy error.

class Model(torch.nn.Module):
    ...
    def forward(self, x):
        # x is an input that requires gradient calculation. x.requires_grad = True
        y = x + 1

        # y is a non-leaf tensor. After assigning this to a specific attribute of the model, deepcopying the model will result in an error.
        self.yy = y

Solution: Check if there are any non-leaf tensors in the model or after the forward pass. Before saving these tensors, make sure to perform the detach operation.

class Model(torch.nn.Module):
    ...
    def forward(self, x):
        # x is an input that requires gradient calculation. x.requires_grad = True
        y = x + 1

        # y is a non-leaf tensor. Do detach operation before saving it in the model.
        self.yy = y.detach()

#Error occurred during export: IndexError: list index out of range.

This error usually occurs because there are qtensor in the model's output.

Solution: Check the model output to ensure that all outputs have been dequantized.