Commits

Branch selector

 jspark1105-patch-3

User selector

All users

Datepicker

All time

Commit History

Commits on Mar 21, 2020

• Update Utils.h

  jspark1105

  authored
  0945f90

• Update RefImplementations.cc

  jspark1105

  authored
  db31580

• Update CMakeLists.txt

  jspark1105

  authored
  6e21cf4

• Quantize should work even if the value after quantization becomes inf (pytorch#317)

  Show description for 3ee0b7d
  

  dskhudia

  authored and

  jspark1105

  committed
  3ee0b7d

• Revert D16234008: Quantize should work even if the value after quantization becomes inf

  Show description for ab431ad
  

  ezyang

  authored and

  jspark1105

  committed
  ab431ad

• Quantize should work even if the value after quantization becomes inf (pytorch#107)

  Show description for 2a50f19
  

  dskhudia

  authored and

  jspark1105

  committed
  2a50f19

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

  Show description for 40235b3
  

  dskhudia

  authored and

  jspark1105

  committed
  40235b3

• Update CODE_OF_CONDUCT.md (pytorch#310)

  Show description for 5562540
  

  jspisak

  authored and

  jspark1105

  committed
  5562540

• Fix the error in calling getCodeLoggingFile() (pytorch#311)

  Show description for d1088e7
  

  jiecaoyu

  authored and

  jspark1105

  committed
  d1088e7

• 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

• fbgemm handles block size 1 sparse adagrad (pytorch#306)

  Show description for e2950ea
  jspark1105

  committed
  e2950ea

• enable parallel for quantization (pytorch#305)

  Show description for f570274
  

  Hongzhang Shan

  authored and

  jspark1105

  committed
  f570274

• Output formatting of results

  Show description for 410b1a0
  

  dskhudia

  authored and

  jspark1105

  committed
  410b1a0

• add copyright to EmbeddingSpMDMTestUtils.{cc,h} (pytorch#303)

  Show description for ca1e41c
  jspark1105

  committed
  ca1e41c

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

  Show description for 1a7be35
  jspark1105

  committed
  1a7be35

• clean up embedding spmdm tests (pytorch#300)

  Show description for 6a4ed1c
  jspark1105

  committed
  6a4ed1c

• Install PDBs for MSVC debug builds (pytorch#301)

  Show description for 536c38c
  

  peterjc123

  authored and

  jspark1105

  committed
  536c38c

• make order btw div and mul in adgrad consistent (#32974)

  Show description for 86dd3bf
  jspark1105

  committed
  86dd3bf

• rowwise sparse adagrad fused (pytorch#298)

  Show description for cfda552
  jspark1105

  committed
  cfda552

• simplify avx2 mask usage in embedding kernels (pytorch#287)

  Show description for ccd91d6
  jspark1105

  committed
  ccd91d6

• change fbgemm export directory from bin to lib (pytorch#296)

  Show description for 496a57f
  

  Hongzhang Shan

  authored and

  jspark1105

  committed
  496a57f

• export more classes for pytorch (pytorch#295)

  Show description for 752a094
  

  Hongzhang Shan

  authored and

  jspark1105

  committed
  752a094

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

  Show description for 878e132
  

  Hongzhang Shan

  authored and

  jspark1105

  committed
  878e132

• Reuse the partition scheme for GEMM in Depthwise Convolution (pytorch#293)

  Show description for 8a77433
  

  jianyuh

  authored and

  jspark1105

  committed
  8a77433

• More fixes to enable USE_FBGEMM for PyTorch on Windows (pytorch#291)

  Show description for d6fa470
  

  Hongzhang Shan

  authored and

  jspark1105

  committed
  d6fa470

• clean up temporary code for isa-dependent serialized packed weight (pytorch#290)

  Show description for 365633b
  jspark1105

  committed
  365633b

• row-wise sparse fp32/fp16/int8 EmbeddingSpMDM (pytorch#288)

  Show description for b8297c2
  jspark1105

  committed
  b8297c2

• enable github action build/test on windows (pytorch#285)

  Show description for b153dbf
  

  Hongzhang Shan

  authored and

  jspark1105

  committed
  b153dbf

• changes for shared release on windows (pytorch#284)

  Show description for d1ebf4e
  

  Hongzhang Shan

  authored and

  jspark1105

  committed
  d1ebf4e

• enable positional weight for EmbeddingSpMDMNBitRowWiseSparse (pytorch#281)

  Show description for fb7a373
  jspark1105

  committed
  fb7a373

• enable EmbeddingSpMDMNBitRowWiseSparse avx2 (pytorch#280)

  Show description for 0b0e2d9
  jspark1105

  committed
  0b0e2d9

• Fix vector size setting and min max set for randFill (pytorch#283)

  Show description for 7f9ced2
  

  Hongzhang Shan

  authored and

  jspark1105

  committed
  7f9ced2

• fix "invalid min and max arguments for uniform_int" error (pytorch#282)

  Show description for defaa88
  

  Hongzhang Shan

  authored and

  jspark1105

  committed
  defaa88

• fix overflow in cache flush (pytorch#260)

  Show description for 2a881a1
  jspark1105

  committed
  2a881a1

Pagination

PreviousNext