Commits

Breadcrumbs

History for

ondevelop

User selector

All users

All time

Commit History

Commits on Sep 29, 2022

• [Dygraph] Support sharding stage2/3+dp in GPT-3 model (#2471)

  Show description for bc23b8b
  

  haohongxiang

  and

  gongenlei

  authored
  bc23b8b

  Copy full SHA for bc23b8b

• [TIPC] Add scripts for npu and xpu, test=develop (#3377)

  Show description for 3abb358
  YanhuiDua

  authored
  3abb358

  Copy full SHA for 3abb358

Commits on Sep 26, 2022

• [MoE] Fix distributed wait api (#3365)

  sljlp

  authored
  07e5c55

  Copy full SHA for 07e5c55

Commits on Sep 21, 2022

• [MoE] Fix recompute & communication api (#3338)

  Show description for 8dc205c
  sljlp

  authored
  8dc205c

  Copy full SHA for 8dc205c

Commits on Sep 14, 2022

• [CodeStyle] Add copyright for python file. (#3259)

  Show description for 815ddfb
  ZHUI

  authored
  815ddfb

  Copy full SHA for 815ddfb

Commits on Sep 7, 2022

• [Recompute] Update recompute for hybrid parallel interface. (#3211)

  Show description for 0ed8de7
  

  wuhuachaocoding

  and

  ZHUI

  authored
  0ed8de7

  Copy full SHA for 0ed8de7

Commits on Sep 6, 2022

• [GPT-3/MOE] Adapt recompute for latest paddle (#3191)

  sljlp

  authored
  b72aa7e

  Copy full SHA for b72aa7e

Commits on Aug 30, 2022

• [MoE] Fix changes of _C_ops usage for paddle develop version. (#3163)

  sljlp

  authored
  a087e0c

  Copy full SHA for a087e0c

Commits on Aug 25, 2022

• add t5 zeroshot demo and add new weights (#3116)

  Show description for e549b57
  

  JunnYu

  and

  guoshengCS

  authored
  e549b57

  Copy full SHA for e549b57

Commits on Aug 22, 2022

• [GPT] Update default lr of args.py (#3101)

  ZHUI

  authored
  bfc8700

  Copy full SHA for bfc8700

Commits on Aug 17, 2022

• [MoE] Fix inner final_state_adamw api usage.

  sljlp

  authored
  e05adbf

  Copy full SHA for e05adbf

Commits on Aug 2, 2022

• [MoE] Fix fluid api bugs for paddle 2.3 (#2937)

  haohongxiang

  authored
  6b4b62d

  Copy full SHA for 6b4b62d

Commits on Jul 21, 2022

• [GPT] Optimize dataset_utils C++ compile (#2846)

  GuoxiaWang

  authored
  4d18e4d

  Copy full SHA for 4d18e4d

Commits on Jul 18, 2022

• [GPT-3] Optimize gpt-3 with fused op (#2650)

  Show description for fe9d03c
  FeixLiu

  authored
  fe9d03c

  Copy full SHA for fe9d03c

Commits on Jul 7, 2022

• [Dy2Stat] Adapt to while transformer refactor logic (#2756)

  Aurelius84

  authored
  00f05b8

  Copy full SHA for 00f05b8

Commits on Jul 6, 2022

• update ernie model to 3.0 (#2728)

  LiuChiachi

  authored
  f923142

  Copy full SHA for f923142

Commits on Jun 30, 2022

• [Benchmark] Adapted for benchmark ips for MOE (#2700)

  Show description for 5e3a588
  sljlp

  authored
  5e3a588

  Copy full SHA for 5e3a588

Commits on Jun 29, 2022

• [MOE] fix bugs of PR-2633 (#2683)

  haohongxiang

  authored
  4d40a2e

  Copy full SHA for 4d40a2e

Commits on Jun 27, 2022

• [CodeStyle] Remove some code with paddle.fluid (#2633)

  Show description for 40417bf
  ZHUI

  authored
  40417bf

  Copy full SHA for 40417bf

Commits on Jun 23, 2022

• inference support amp (#2505)

  Show description for 8dd2e2a
  

  LiYuRio

  and

  ZeyuChen

  authored
  8dd2e2a

  Copy full SHA for 8dd2e2a

• [GPT-3] Fix reader of gpt-3 static graph for pipeline mode (#2564)

  sljlp

  authored
  634c2e7

  Copy full SHA for 634c2e7

Commits on Jun 16, 2022

• Update fluid APIs (#2469)

  Show description for f10575d
  FrostML

  authored
  f10575d

  Copy full SHA for f10575d

• [NEW MODEL] Add_XLM_model (#2080)

  Show description for 07bafdc
  

  JunnYu

  and

  yingyibiao

  authored
  07bafdc

  Copy full SHA for 07bafdc

Commits on Jun 14, 2022

• [GPT-3] Add GPT configs 89B and 175B (#2504)

  Show description for 1f446ff
  LiYuRio

  authored
  1f446ff

  Copy full SHA for 1f446ff

Commits on Jun 13, 2022

• [GPT-3] Add fused_multi_transformer for inference (#2500)

  wangxicoding

  authored
  016db3c

  Copy full SHA for 016db3c

• update yapf (#2464)

  FrostML

  authored
  7c06b02

  Copy full SHA for 7c06b02

Commits on Jun 9, 2022

• [MoE] Support Eager Dygraph and Group Sharded Parallel in MoE model (#2392)

  Show description for 636e2e5
  haohongxiang

  authored
  636e2e5

  Copy full SHA for 636e2e5

• Add MOE benchmark (#2404)

  Show description for e4de45f
  sljlp

  authored
  e4de45f

  Copy full SHA for e4de45f

Commits on Jun 7, 2022

• delete old style code for ernie hybrid parallel. (#2429)

  ZHUI

  authored
  cec1caf

  Copy full SHA for cec1caf

Commits on Jun 6, 2022

• Fix datas to data (#2410)

  Show description for 75edf33
  yeliang2258

  authored
  75edf33

  Copy full SHA for 75edf33

Commits on Jun 2, 2022

• [MoE] Add MoE model pretrain and support MoE+DP (#2385)

  Show description for 75735a8
  haohongxiang

  authored
  75735a8

  Copy full SHA for 75735a8

Commits on Jun 1, 2022

• Delete pretraining_data_prepare of bert (#2379)

  Show description for 25b41ea
  ZHUI

  authored
  25b41ea

  Copy full SHA for 25b41ea

Commits on May 31, 2022

• gpt fused attention and feedforward (#2277)

  Show description for 1a8c099
  wangxicoding

  authored
  1a8c099

  Copy full SHA for 1a8c099

Commits on May 26, 2022

• [GPT-3][dygraph] Fix pipeline in new dygraph (#2314)

  Show description for 7c3a7e4
  

  ForFishes

  and

  ZHUI

  authored
  7c3a7e4

  Copy full SHA for 7c3a7e4

Commits on May 20, 2022

• [GPT-3] Add gpt-3 dist inference demo scripts. (#2265)

  wangxicoding

  authored
  ee5aa26

  Copy full SHA for ee5aa26