[nnc] Fix half2float conversion and re-enable float16 (#64199)
authorBert Maher <bertrand@fb.com>
Tue, 31 Aug 2021 01:36:33 +0000 (18:36 -0700)
committerFacebook GitHub Bot <facebook-github-bot@users.noreply.github.com>
Tue, 31 Aug 2021 01:37:55 +0000 (18:37 -0700)
commitebc0aacf83a0446ed798a96059c05da815c73d3d
treefdbde7205a716ac7d31b3cf73a020bc66628118c
parent1f16c22dc8251f01627ee73ad1ef69bd18e51447
[nnc] Fix half2float conversion and re-enable float16 (#64199)

Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/64199

Test Plan: Imported from OSS

Reviewed By: ZolotukhinM

Differential Revision: D30643865

Pulled By: bertmaher

fbshipit-source-id: 9de6adca53bd08839328cbaf6364f7de9550264b
test/test_jit_fuser_te.py
test/test_tensorexpr.py
torch/csrc/jit/passes/tensorexpr_fuser.cpp
torch/csrc/jit/tensorexpr/half_support.h
torch/csrc/jit/tensorexpr/ir_verifier.cpp