Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Sep 15, 2020

• Merge pull request NVIDIA#959 from a-maci/update-ASP-readme

  Show description for 4a1fa2c
  thorjohnsen

  authored
  4a1fa2c

  Copy full SHA for 4a1fa2c

Commits on Sep 14, 2020

• Merge pull request NVIDIA#5 from a-maci/a-maci-patch-update-asp-readme

  Show description for 48fc613
  a-maci

  authored
  48fc613

  Copy full SHA for 48fc613

• Update README for ASP

  Show description for e3794f4
  a-maci

  authored
  e3794f4

  Copy full SHA for e3794f4

Commits on Aug 15, 2020

• Should pass stricter stride/size checks in pytorch (NVIDIA#942)

  mcarilli

  authored
  4ef930c

  Copy full SHA for 4ef930c

Commits on Aug 10, 2020

• move sm80 code inside MHA (NVIDIA#937)

  Show description for 5d9b5cb
  ptrblck

  authored
  5d9b5cb

  Copy full SHA for 5d9b5cb

Commits on Aug 6, 2020

• sgd supports zero-grad (NVIDIA#926)

  

  ngimel

  authored
  85b1783

  Copy full SHA for 85b1783

Commits on Aug 5, 2020

• set device guard for multi tensor optimizer implementations (NVIDIA#927)

  Show description for 274cc06
  

  ngimel

  authored
  274cc06

  Copy full SHA for 274cc06

Commits on Aug 1, 2020

• Add sm80 for CUDA >= 11 (NVIDIA#925)

  ptrblck

  authored
  5b53121

  Copy full SHA for 5b53121

Commits on Jul 30, 2020

• fix default mode missing additive mask option (NVIDIA#924)

  Show description for 700d682
  

  seryilmaz

  and

  Sukru Eryilmaz

  authored
  700d682

  Copy full SHA for 700d682

Commits on Jul 23, 2020

• Merge pull request NVIDIA#918 from a-maci/ASP_sparse_param_dict_update

  Show description for 459de22
  thorjohnsen

  authored
  459de22

  Copy full SHA for 459de22

Commits on Jul 22, 2020

• Merge pull request NVIDIA#4 from a-maci/ASP_sparse_param_dict_update

  Show description for eb5e96c
  a-maci

  authored
  eb5e96c

  Copy full SHA for eb5e96c

• Accept custom (layer type:param name) to include in sparse_parameter dictionary

  Show description for b3c1641
  a-maci

  authored
  b3c1641

  Copy full SHA for b3c1641

• Merge pull request NVIDIA#3 from NVIDIA/master

  Show description for eb95950
  a-maci

  authored
  eb95950

  Copy full SHA for eb95950

Commits on Jul 21, 2020

• Merge pull request NVIDIA#917 from a-maci/master

  Show description for 0ac5dd6
  thorjohnsen

  authored
  0ac5dd6

  Copy full SHA for 0ac5dd6

Commits on Jul 20, 2020

• Merge pull request NVIDIA#2 from a-maci/a-maci-patch-1

  Show description for 089149d
  a-maci

  authored
  089149d

  Copy full SHA for 089149d

• Fixing mask multiplication with grad tensors

  Show description for 774de91
  a-maci

  authored
  774de91

  Copy full SHA for 774de91

• Merge pull request NVIDIA#1 from NVIDIA/master

  Show description for 3dd3607
  a-maci

  authored
  3dd3607

  Copy full SHA for 3dd3607

Commits on Jul 16, 2020

• Merge pull request NVIDIA#910 from szmigacz/smigacz/mha_xavier_init_gain_fix

  Show description for 3104fd5
  thorjohnsen

  authored
  3104fd5

  Copy full SHA for 3104fd5

• Merge pull request NVIDIA#904 from ksivaman/sparsity

  Show description for 4027bcb
  thorjohnsen

  authored
  4027bcb

  Copy full SHA for 4027bcb

Commits on Jul 9, 2020

• Fixed weight init for fused weight matrices in fused MHA by adding correct gain factor.

  szmigacz

  committed
  a0d99fd

  Copy full SHA for a0d99fd

Commits on Jul 6, 2020

• [sync BN] (NVIDIA#792)

  Show description for 1ff54b8
  jjsjann123

  authored
  1ff54b8

  Copy full SHA for 1ff54b8

Commits on Jul 1, 2020

• name -> p_name (name variable is out of scope)

  

  Kirthi Sivamani

  committed
  59995c7

  Copy full SHA for 59995c7

Commits on Jun 30, 2020

• Don't patch tensor ops that aren't present (NVIDIA#899)

  Show description for 43a6f9f
  

  mcarilli

  and

  definitelynotmcarilli

  authored
  43a6f9f

  Copy full SHA for 43a6f9f

Commits on Jun 23, 2020

• Merge pull request NVIDIA#892 from kexinyu/master

  Show description for 44532b3
  kexinyu

  authored
  44532b3

  Copy full SHA for 44532b3

• add test case for non-zero weight decay

  kexinyu

  committed
  ad50ce9

  Copy full SHA for ad50ce9

• test nvlamb; hyperparams consistent with adam/adagrad tests

  kexinyu

  committed
  cd3d6d1

  Copy full SHA for cd3d6d1

• add test for FusedLAMB

  kexinyu

  committed
  9774ce0

  Copy full SHA for 9774ce0

Commits on Jun 15, 2020

• Merge pull request NVIDIA#885 from a-maci/2dmasking_sparsity

  Show description for c3fad1a
  thorjohnsen

  authored
  c3fad1a

  Copy full SHA for c3fad1a

• Updating comment

  Show description for 73ff00e
  a-maci

  authored
  73ff00e

  Copy full SHA for 73ff00e

• adding comments for 2d pruning

  Show description for 8923046
  a-maci

  authored
  8923046

  Copy full SHA for 8923046

• editing comments

  a-maci

  authored
  ceca097

  Copy full SHA for ceca097

Commits on Jun 11, 2020

• Merge pull request NVIDIA#883 from NVIDIA/schetlur/stream_bug_fix

  Show description for 02a3387
  schetlur

  authored
  02a3387

  Copy full SHA for 02a3387

• Update softmax.h

  Show description for 060bd5c
  schetlur

  authored
  060bd5c

  Copy full SHA for 060bd5c

Commits on Jun 10, 2020

• Merge pull request NVIDIA#880 from seryilmaz/seryilmaz/stream

  Show description for 058addb
  cliffwoolley

  authored
  058addb

  Copy full SHA for 058addb

• add streamid

  

  Sukru Eryilmaz

  committed
  1574c03

  Copy full SHA for 1574c03

Pagination

Previous

Next