projects / platform / upstream / pytorch.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 078b800) | patch
[fx2trt] add unsqueeze converter (#63355)
authorShiyan Deng <dsy842974287@fb.com>
Mon, 16 Aug 2021 22:16:51 +0000 (15:16 -0700)
committerFacebook GitHub Bot <facebook-github-bot@users.noreply.github.com>
Mon, 16 Aug 2021 22:18:43 +0000 (15:18 -0700)
commitf4f2c1231a4a429182d12b58e16b4a42fcae68a5
treee7def2f151203d78ed26ff1e6b006bea93b044fetree | snapshot
parent078b8004a62a51f75e1fbd8d08eea359af6bb1d7commit | diff
[fx2trt] add unsqueeze converter (#63355)

Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/63355

Added converter for acc_ops.unsqueeze. Needed for ig model.

DIdn't add support for input that has more than one dynamic dim. This is not needed right now and I feel it would be a rare case.

Test Plan: unit test

Reviewed By: yinghai

Differential Revision: D30138293

fbshipit-source-id: 899fe8eb68387de83195a2f6e199618d96f09a9e
torch/fx/experimental/fx2trt/converters/acc_ops_converters.py diff | blob | history
torch/fx/experimental/fx2trt/fx2trt.py diff | blob | history
Domain: Machine Learning / ML Framework;