Angela Yi
Angela Yi
Summary: https://github.com/pytorch/pytorch/pull/118239 added a decomposition for pixel_shuffle, so pixel_shuffle no longer needs to be a Core ATen Op. We have also fixed the internal use case so that it no...
By changing runtime symbolic asserts to using assert_scalar, the asserts can call into `expect_true` and modify the shape env so that we can run through the traced graph module with...
Reviewed By: tarun292 Differential Revision: D56209661
Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * __->__ #125054 cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @peterbell10 @ipiszy @yf225 @chenyang78 @kadeng @muchulee8 @ColinPeppler @amjames @desertfire @chauhang
Fixes #ISSUE_NUMBER
In the given test case, we have a ModuleList of 3 modules (`norm.0`, `norm.1`, `norm.2`) which share the same `weight` and `bias` tensors. However when we trace, they all end...
Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom): * __->__ #125242 cc @ezyang @msaroufim @bdhirsh @anijain2305 @chauhang @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @peterbell10 @ipiszy @yf225 @chenyang78 @kadeng...
Fixes [internal error](https://fb.workplace.com/groups/1075192433118967/permalink/1416709435633930/). The issue is that the asserting nodes added in the `insert_deferred_runtime_assertion` pass do not contain metadata that the ExportedProgram requires the graph to have. One solution to...
A previous PR added support for serializing/deserializing example inputs, but this fails when `example_inputs` is none.
Test Plan: CI Differential Revision: D56893728 cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @chenyang78 @kadeng @chauhang