shingjan commented on code in PR #14885:
URL: https://github.com/apache/tvm/pull/14885#discussion_r1198195718
##########
python/tvm/relay/frontend/pytorch.py:
##########
@@ -2260,7 +2260,8 @@ def broadcast_tensors(self, inputs, input_types):
tensor_list = inputs[0]
import torch
- res_shape = list(torch.broadcast_shapes(*[self.infer_shape(t) for t in
tensor_list]))
+ infer_shape_value = [self.infer_shape(t) for t in tensor_list]
Review Comment:
If we are trying to be backward-compatible here, would that be possible that
we can check if the `broadcast_shape` available first, and fall back to
`broadcast_tensors.shape` if not so?
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
