Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Feb 17, 2021

• Add unit test for UVM APIs in FBGEMM_GPU (pytorch#516)

  Show description for 887359c
  

  Joe Hu

  authored and

  facebook-github-bot

  committed
  887359c

• Cooperlake performance debug: fix partition issues (pytorch#514)

  Show description for c520088
  

  jiyuanzFB

  authored and

  facebook-github-bot

  committed
  c520088

Commits on Feb 16, 2021

• Replaced APPROX_ADAGRAD and APPROX_ROWWISE_ADAGRAD with EXACT_ADAGRAD and EXACT_ROWWISE_ADAGRAD

  Show description for aa1b4e6
  

  Andrew Wang

  authored and

  facebook-github-bot

  committed
  aa1b4e6

• Fix 10x model

  Show description for b51f3e6
  

  Dmytro Ivchenko

  authored and

  facebook-github-bot

  committed
  b51f3e6

• Fix/Optimize the thread block allocation in sparse_ops (pytorch#494)

  Show description for 24e92f0
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  24e92f0

Commits on Feb 13, 2021

• remove temp interface for sparse adagrad (pytorch#515)

  Show description for 7f3baec
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  7f3baec

Commits on Feb 11, 2021

• add int8 table batched embedding ops bench

  Show description for 8db2382
  

  amylittleyang

  authored and

  facebook-github-bot

  committed
  8db2382

• remove old sparse adagrad interface (pytorch#513)

  Show description for 58497f7
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  58497f7

• enable uint8 embedding

  Show description for 4d20325
  

  amylittleyang

  authored and

  facebook-github-bot

  committed
  4d20325

Commits on Feb 10, 2021

• clang format BenchUtils.h (pytorch#512)

  Show description for 884fb25
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  884fb25

Commits on Feb 9, 2021

• Adaptive Weight decay based on Sparse feature ID frequency. (pytorch#508)

  Show description for 092497f
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  092497f

• remove warning unused variable remainder (pytorch#511)

  Show description for a8bee90
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  a8bee90

Commits on Feb 8, 2021

• removing parallel build for OSS (pytorch#509)

  Show description for c5e2273
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  c5e2273

Commits on Feb 6, 2021

• Back out "fix thread block size issue for cooperlake"

  Show description for 43570c8
  

  jiecaoyu

  authored and

  facebook-github-bot

  committed
  43570c8

Commits on Feb 5, 2021

• Enable fbgemm TQN

  Show description for fc29382
  

  Xiaohan Wei

  authored and

  facebook-github-bot

  committed
  fc29382

Commits on Feb 4, 2021

• fix thread block size issue for cooperlake (pytorch#504)

  Show description for 4ffd475
  

  jiyuanzFB

  authored and

  facebook-github-bot

  committed
  4ffd475

• Add emb/cache precision supports in HPC PT trainer (pytorch#507)

  Show description for b8e2c23
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  b8e2c23

• Fix typo (pytorch#506)

  Show description for fe1fe6d
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  fe1fe6d

Commits on Feb 3, 2021

• torch.package support

  Show description for a4de38e
  

  Dmytro Ivchenko

  authored and

  facebook-github-bot

  committed
  a4de38e

Commits on Feb 2, 2021

• Enhance Split Table Batched Embedding prefetch code.

  Show description for d0222fd
  

  another-pjohnson

  authored and

  facebook-github-bot

  committed
  d0222fd

Commits on Feb 1, 2021

• Revert the changes on reducing the thread block number in BWD (pytorch#503)

  Show description for 1a8f2ea
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  1a8f2ea

• Add special handler on is_uvm_tensor checker for CPU tensor (pytorch#502)

  Show description for 8d0b883
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  8d0b883

• Use the latest philox_cuda_state API for stochastic rounding (#51004)

  Show description for d244512
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  d244512

Commits on Jan 31, 2021

• Reduce the number of launched threads in the backward pass (pytorch#501)

  Show description for 1962324
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  1962324

Commits on Jan 29, 2021

• Add instructions for generating documents in OSS FBGEMM README.md file (pytorch#499)

  Show description for a74e767
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  a74e767

Commits on Jan 27, 2021

• Fix the unit test of test_cache_pipeline (pytorch#498)

  Show description for bbb04cd
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  bbb04cd

• add comment on do_clip (pytorch#495)

  Show description for e345a99
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  e345a99

Commits on Jan 26, 2021

• Add stochastic rounding option in the op-level benchmark for split table batched embeddings (pytorch#496)

  Show description for 27dfaf9
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  27dfaf9

Commits on Jan 25, 2021

• Only keep one copy of split_table_batched_embeddings_ops.py (pytorch#492)

  Show description for 1ac1339
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  1ac1339

Commits on Jan 22, 2021

• support row-wise uint8 quantization in split embedding kernels (pytorch#491)

  Show description for 29fa98e
  

  amylittleyang

  authored and

  facebook-github-bot

  committed
  29fa98e

• fix stochastic rounding

  Show description for ceb16c9
  

  Yinbin Ma

  authored and

  facebook-github-bot

  committed
  ceb16c9

• Update a few things for OSS (pytorch#490)

  Show description for adbd4d8
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  adbd4d8

Commits on Jan 21, 2021

• Use memory read kernel to flush the cache (pytorch#487)

  Show description for 272c6c3
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  272c6c3

Commits on Jan 20, 2021

• Update README.md (pytorch#489)

  Show description for 0d34840
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  0d34840

• Detect ISA once, not repeatedly (pytorch#480)

  Show description for 2799d8f
  

  swolchok

  authored and

  facebook-github-bot

  committed
  2799d8f

Pagination

PreviousNext