projects / platform / upstream / pytorch.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 9d95d48) | patch
Adding BFP16 quantization/dequantization support to OSS (#63059)
authorMarjan Fariborz <marjanf@fb.com>
Thu, 26 Aug 2021 06:40:09 +0000 (23:40 -0700)
committerFacebook GitHub Bot <facebook-github-bot@users.noreply.github.com>
Thu, 26 Aug 2021 06:41:34 +0000 (23:41 -0700)
commit3b284ab0243d22bb831a1685f47061b9612e1cb3
tree5a0577adc86f572ebb9012fadac90bb1ce68041etree | snapshot
parent9d95d485679392774532d4c79a73b9c11b665e1bcommit | diff
Adding BFP16 quantization/dequantization support to OSS (#63059)

Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/63059

Supporting BFP16 quantization method to OSS. Currently only support CPU
ghstack-source-id: 136639528

Test Plan: Imported from OSS

Reviewed By: wanchaol

Differential Revision: D30194538

fbshipit-source-id: ac248567ad8028457c2a91b77ef2ce81709fce53
test/distributed/algorithms/quantization/test_quantization.py diff | blob | history
tools/build_variables.bzl diff | blob | history
torch/csrc/distributed/c10d/init.cpp diff | blob | history
torch/csrc/distributed/c10d/quantization/quantization.cpp [new file with mode: 0644] blob
torch/csrc/distributed/c10d/quantization/quantization.h [new file with mode: 0644] blob
torch/csrc/distributed/c10d/quantization/quantization_gpu.cu [new file with mode: 0644] blob
torch/csrc/distributed/c10d/quantization/quantization_gpu.h [new file with mode: 0644] blob
torch/csrc/distributed/c10d/quantization/quantization_utils.h [new file with mode: 0644] blob
torch/distributed/algorithms/quantization/quantization.py diff | blob | history
Domain: Machine Learning / ML Framework;