Commits

Branch selector

 main

User selector

jspark1105

Datepicker

All time

Commit History

Commits on Jun 28, 2024

• add functional CPU VBE support (pytorch#2773)

  Show description for 565ab30
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  565ab30

Commits on Feb 25, 2024

• bug fix for CPU pruned weighted TBE

  Show description for 2bba5cc
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  2bba5cc

Commits on Apr 4, 2023

• reduce_to_one (pytorch#1571)

  Show description for 595adad
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  595adad

Commits on Jan 31, 2023

• fix comment related num SMs in H100 (pytorch#1569)

  Show description for b01cc64
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  b01cc64

Commits on Jan 30, 2023

• bug fix atomics without fence and speedup super long segments (pytorch#1563)

  Show description for b2034ae
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  b2034ae

Commits on Jan 1, 2023

• support jagged tensor with no inner dense dim (pytorch#1267)

  Show description for 32b6a12
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  32b6a12

• use -1 for sentinel value for direct mapped cache (pytorch#1520)

  Show description for baa5229
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  baa5229

Commits on Dec 23, 2022

• increase deadline for test_quantized_comm_codec (pytorch#1521)

  Show description for 79358d7
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  79358d7

Commits on Nov 15, 2022

• add Meta and Autograd backends for jagged ops (pytorch#1461)

  Show description for d47449d
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  d47449d

Commits on Nov 8, 2022

• handle case when value size is bigger than last value of offset (pytorch#1454)

  Show description for 0fae3a3
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  0fae3a3

Commits on Oct 12, 2022

• use anonymous namespace in split_embeddings_cache_cuda.cu when appropriate (pytorch#1389)

  Show description for 6799342
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  6799342

• simplify direct mapped cache parallelization (pytorch#1386)

  Show description for d922e24
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  d922e24

• add unit test for direct mapped cache (pytorch#1385)

  Show description for bdf9090
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  bdf9090

Commits on Oct 10, 2022

• pyre fixes (pytorch#1384)

  Show description for a2186e1
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  a2186e1

• copyright headers (pytorch#1387)

  Show description for 69f6075
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  69f6075

Commits on Sep 28, 2022

• masked_select_jagged_1d (pytorch#1361)

  Show description for 102655d
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  102655d

Commits on Sep 24, 2022

• truncate_jagged_1d (pytorch#1345)

  Show description for 6da1b24
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  6da1b24

Commits on Sep 14, 2022

• multiple thread blocks + atomics for very long runs in emb bwd (pytorch#1283)

  Show description for 2be247c
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  2be247c

Commits on Sep 13, 2022

• refactor emb bwd cta_per_row (pytorch#1282)

  Show description for 2d822dc
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  2d822dc

• fewer threads per row in emb bwd cta_per_row for small D (pytorch#1279)

  Show description for 1c47b74
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  1c47b74

Commits on Sep 11, 2022

• use fewer threads per row in emb bwd warp_per_row for small D (pytorch#1278)

  Show description for 4694251
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  4694251

• use fewer threads per row in emb fwd for small D (pytorch#1277)

  Show description for 411e550
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  411e550

Commits on Sep 1, 2022

• cuda zipf generation from cupy (pytorch#1272)

  Show description for 2078e87
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  2078e87

Commits on Aug 30, 2022

• use fewer thread blocks for kernels in split_embeddings_cache_cuda.cu (pytorch#1268)

  Show description for 396217b
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  396217b

• use fewer thread blocks for lru/lfu_cache_find_uncached_kernel (pytorch#1265)

  Show description for a51cbe8
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  a51cbe8

Commits on Jul 4, 2022

• corner case handling of empty values tensor in jagged_to_padded_dense CPU op (pytorch#1187)

  Show description for ffc05b3
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  ffc05b3

Commits on Jun 23, 2022

• zipf w/o replacement faster in split_table_batched_embeddings_benchmark.py (pytorch#1170)

  Show description for 8e9b829
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  8e9b829

Commits on Jun 22, 2022

• move jagged tensor tests to a separate file (pytorch#1171)

  Show description for 38993ce
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  38993ce

Commits on May 27, 2022

• BatchedHyperCompressedSparseColumn -> HyperCompressedSparseColumn (pytorch#1137)

  Show description for 99a6a5f
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  99a6a5f

Commits on Apr 27, 2022

• add bf16 support in jagged tensor ops (pytorch#1079)

  Show description for 4454ac5
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  4454ac5

Commits on Apr 20, 2022

• run test_batched_dense_vec_jagged_2d_mul on CPU (pytorch#1062)

  Show description for 40d85ff
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  40d85ff

Commits on Apr 19, 2022

• test permute102_baddbmm_permute102 with float (pytorch#1059)

  Show description for 80c0268
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  80c0268

Commits on Apr 15, 2022

• add 7x7 depthwise (pytorch#1049)

  Show description for 40101e8
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  40101e8

Commits on Apr 14, 2022

• optimize jagged_dense_elementwise_jagged_output (pytorch#1046)

  Show description for 61de99f
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  61de99f

Commits on Mar 29, 2022

• handle truncation vs. padding case in jagged tensors (pytorch#1013)

  Show description for a0ff6de
  

  jspark1105

  authored and

  facebook-github-bot

  committed
  a0ff6de

Pagination

PreviousNext