shiwenloong opened a new pull request #6226: URL: https://github.com/apache/incubator-tvm/pull/6226
Std op in torchscript supports overloading (https://pytorch.org/docs/stable/jit_builtin_functions.html#builtin-functions) : ``` torch.std(self : Tensor, unbiased : bool=True) -> Tensor torch.std(self : Tensor, dim : List[int], unbiased : bool=True, keepdim : bool=False) -> Tensor ``` The std op without specified dimensions was not supported in current pytorch frontend, so pytorch module like below can't be converted. ``` class StdModule(nn.Module): def forward(self, *args): return args[0].std(unbiased=False) ``` This PR fixes the problem. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected]
