Commits

Breadcrumbs

History for

one1abaec

User selector

All users

Datepicker

All time

Commit History

Commits on Mar 10, 2021

• print total execution time of mkl (pytorch#537)

  Show description for d12fc48
  

  YangWang92

  authored and

  facebook-github-bot

  committed
  d12fc48

Commits on Feb 27, 2021

• open source sparse-dense matrix-matrix multiplication (int8 and fp32) (pytorch#523)

  Show description for 953e671
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  953e671

Commits on Feb 18, 2021

• Update SparseAdagradBenchmark with adjustable weight decay (pytorch#517)

  Show description for 975e41d
  

  venkatacrc

  authored and

  facebook-github-bot

  committed
  975e41d

Commits on Feb 11, 2021

• remove old sparse adagrad interface (pytorch#513)

  Show description for 58497f7
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  58497f7

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 Nov 10, 2020

• Remove dead includes in deeplearning/fbgemm/bench (pytorch#456)

  Show description for 92c5f37
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  92c5f37

Commits on Nov 9, 2020

• Generalize GEMM Benchmark (pytorch#447)

  Show description for eb55572
  

  efiks

  authored and

  facebook-github-bot

  committed
  eb55572

Commits on Oct 20, 2020

• add im2col for transposed conv path (pytorch#445)

  Show description for 23cb1db
  

  YazhiGao

  authored and

  facebook-github-bot

  committed
  23cb1db

Commits on Oct 9, 2020

• match old sparse adagrad interface with the new one (pytorch#434)

  Show description for 75ea7ce
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  75ea7ce

• add new sparse adagrad interface takes weight_decay at call time not at codegen time (pytorch#432)

  Show description for 974d2b4
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  974d2b4

Commits on Oct 2, 2020

• Matrix transpose kernels for int8/uint8 datatypes (pytorch#431)

  Show description for fe91640
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  fe91640

Commits on Sep 4, 2020

• part2: Move embedding quantization kernels to fbgemm for better sharing between C2/PT (pytorch#425)

  Show description for 7e401c0
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  7e401c0

Commits on Aug 31, 2020

• Move embedding quantization kernels to fbgemm for better sharing between C2/PT (pytorch#419)

  Show description for a9397bf
  

  dskhudia

  authored and

  facebook-github-bot

  committed
  a9397bf

Commits on Aug 8, 2020

• use new more general depthwise 3d conv interface (#42697)

  Show description for fbd813e
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  fbd813e

Commits on Aug 7, 2020

• enable 3x5x5 and 5x5x5 3d depthwise conv (pytorch#399)

  Show description for 0b01501
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  0b01501

Commits on Jun 1, 2020

• fix benchmark to consider offsets is now default instead of lengths (pytorch#383)

  Show description for 377c316
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  377c316

Commits on May 15, 2020

• Add conv_1d (pytorch#369)

  Show description for 17b31be
  

  Hongzhang Shan

  authored and

  facebook-github-bot

  committed
  17b31be

Commits on May 14, 2020

• dirsync fbgemm with xplat

  Show description for 7ed5f9f
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  7ed5f9f

Commits on May 13, 2020

• fix benchmark to consider offsets is now default instead of lengths (pytorch#373)

  Show description for 483ccc9
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  483ccc9

Commits on May 6, 2020

• unified conv to call dw conv with 2 oc per g (pytorch#360)

  Show description for 40f530d
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  40f530d

• depthwise convolution with 2 output channels per group (pytorch#359)

  Show description for dfbb13f
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  dfbb13f

Commits on May 1, 2020

• remove one-stage interface for sparse adagrad (pytorch#366)

  Show description for b59c633
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  b59c633

Commits on Apr 10, 2020

• Move FakeFP16 back to internal to remove dependency on MKL (#36297)

  Show description for 20b6cf1
  

  jianyuh

  authored and

  facebook-github-bot

  committed
  20b6cf1

Commits on Apr 9, 2020

• Open sourcing fbgemm_fp16 ops (#36212)

  Show description for 1910f8c
  

  Yinghai Lu

  authored and

  facebook-github-bot

  committed
  1910f8c

Commits on Apr 3, 2020

• add dummy compute to cache evict func (pytorch#337)

  Show description for 1bda6ae
  jspark1105

  committed
  1bda6ae

Commits on Mar 27, 2020

• Rename times global to avoid conflict

  Show description for a6e81fb
  

  andrewjcg

  authored and

  facebook-github-bot

  committed
  a6e81fb

Commits on Mar 23, 2020

• minor fixes in embedding benchmarks (pytorch#326)

  Show description for 49a56f7
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  49a56f7

Commits on Mar 22, 2020

• prefetch scale/bias (pytorch#322)

  Show description for d7c4a34
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  d7c4a34

Commits on Mar 21, 2020

• Make spmmfp32 and spmmi8 benchmark consistent with tests (pytorch#315)

  Show description for 40235b3
  

  dskhudia

  authored and

  jspark1105

  committed
  40235b3

• reduce Windows CI warnings (pytorch#309)

  Show description for 99df751
  jspark1105

  committed
  99df751

• make cmake less verbose (pytorch#308)

  Show description for 4942ea9
  jspark1105

  committed
  4942ea9

• Output formatting of results

  Show description for 410b1a0
  

  dskhudia

  authored and

  jspark1105

  committed
  410b1a0

• fix index mapping table type to int32_t (pytorch#302)

  Show description for 1a7be35
  jspark1105

  committed
  1a7be35

• rowwise sparse adagrad fused (pytorch#298)

  Show description for cfda552
  jspark1105

  committed
  cfda552

• Add ASAN check for github action (pytorch#294)

  Show description for 878e132
  

  Hongzhang Shan

  authored and

  jspark1105

  committed
  878e132

Pagination

PreviousNext