Commits

Breadcrumbs

History for

ondevelop

User selector

All users

All time

Commit History

Commits on Sep 7, 2023

• Add intokens support for chatglm2 (#6941)

  w5688414

  authored
  adee5a5

  Copy full SHA for adee5a5

Commits on Sep 5, 2023

• Add llama alibi intokens support (#6921)

  Show description for 116c71c
  w5688414

  authored
  116c71c

  Copy full SHA for 116c71c

Commits on Aug 29, 2023

• [llm benchmark] Add llama flash attentions & GPT3 (#6603)

  Show description for c5b78ba
  

  w5688414

  and

  sijunhe

  authored
  c5b78ba

  Copy full SHA for c5b78ba

Commits on Aug 14, 2023

• Fix paddle.no_grad in moe example. (#6715)

  guoshengCS

  authored
  09329bd

  Copy full SHA for 09329bd

Commits on Aug 13, 2023

• Implement InTokens data stream for ChatGLM (#6701)

  Show description for 46e5551
  sijunhe

  authored
  46e5551

  Copy full SHA for 46e5551

Commits on Aug 10, 2023

• Fix unwanted fields in InTokens and peft benchmark (#6682)

  Show description for ea89a5d
  sijunhe

  authored
  ea89a5d

  Copy full SHA for ea89a5d

Commits on Aug 8, 2023

• Chatglmv2 attention mask (#6634)

  Show description for c394640
  wtmlon

  authored
  c394640

  Copy full SHA for c394640

• Remove most of fluid usage (#6636)

  Show description for a732077
  guoshengCS

  authored
  a732077

  Copy full SHA for a732077

Commits on Aug 3, 2023

• Add InTokens to llm; Improve InTokens data efficiency (#6552)

  Show description for 1324998
  sijunhe

  authored
  1324998

  Copy full SHA for 1324998

Commits on Aug 1, 2023

• [llm benchmark] Update benchmark (#6529)

  Show description for 018755f
  w5688414

  authored
  018755f

  Copy full SHA for 018755f

Commits on Jul 31, 2023

• revert (#6561)

  sijunhe

  authored
  5ac5803

  Copy full SHA for 5ac5803

Commits on Jul 28, 2023

• [llm benchmark] Update intokens results (#6548)

  Show description for c117c2d
  w5688414

  authored
  c117c2d

  Copy full SHA for c117c2d

Commits on Jul 27, 2023

• [llm benchmark] Add intokens (#6436)

  Show description for 537665a
  w5688414

  authored
  537665a

  Copy full SHA for 537665a

• Fix `LlamaTokenizer`'s `pad_token`; Fix `eos_token_id` logic (#6520)

  Show description for bb63581
  sijunhe

  authored
  bb63581

  Copy full SHA for bb63581

Commits on Jul 24, 2023

• Add llm gpt3 (#6455)

  Show description for 3fd1834
  wawltor

  authored
  3fd1834

  Copy full SHA for 3fd1834

• fix styles (#6472)

  sijunhe

  authored
  d9e1d5f

  Copy full SHA for d9e1d5f

• Fix running model with shard group stage3 (#6466)

  Show description for 588e6b2
  

  pangengzheng

  and

  sijunhe

  authored
  588e6b2

  Copy full SHA for 588e6b2

Commits on Jul 21, 2023

• Add C-Eval scripts (#6453)

  Show description for a7a5d64
  sijunhe

  authored
  a7a5d64

  Copy full SHA for a7a5d64

Commits on Jul 20, 2023

• [LLM] Upgrade sharding stage1 for new develop paddle (#6360)

  liuzhenhai93

  authored
  106ee7c

  Copy full SHA for 106ee7c

• [chatglm] rename chatglm to causallm (#6444)

  Show description for 55efa3d
  lugimzzz

  authored
  55efa3d

  Copy full SHA for 55efa3d

Commits on Jul 19, 2023

• create the llm directory and remove the example llm model to the llm directory (#6426)

  Show description for 125a126
  wawltor

  authored
  125a126

  Copy full SHA for 125a126

• fix the bloom model evaluation (#6400)

  wawltor

  authored
  b5ca5bc

  Copy full SHA for b5ca5bc

• Support ChatGLM2 Training (FT, LoRA and Prefix) (#6402)

  Show description for acfd537
  sijunhe

  authored
  acfd537

  Copy full SHA for acfd537

Commits on Jul 18, 2023

• [LLM] Llama extend RotaryEmbedding context(Adaptive NTK) (#6312)

  Show description for 7c6772b
  ZHUI

  authored
  7c6772b

  Copy full SHA for 7c6772b

Commits on Jul 17, 2023

• support json parse (#6415)

  

  gongenlei

  authored
  5b96cdc

  Copy full SHA for 5b96cdc

Commits on Jul 14, 2023

• change numpy.bool to numpy.bool_ (#6401)

  zhiqiu

  authored
  cfbe700

  Copy full SHA for cfbe700

Commits on Jul 13, 2023

• Support multi-gpu for ChatGLM-v2 (#6373)

  Show description for 583e9c0
  sijunhe

  authored
  583e9c0

  Copy full SHA for 583e9c0

• [llm benchmark] Add qlora benchmark (#6376)

  Show description for 0646056
  w5688414

  authored
  0646056

  Copy full SHA for 0646056

• Add Falcon single-gpu (#6345)

  Show description for 07e5560
  wtmlon

  authored
  07e5560

  Copy full SHA for 07e5560

Commits on Jul 12, 2023

• [llm]fix shift label (#6369)

  Show description for 84af93f
  lugimzzz

  authored
  84af93f

  Copy full SHA for 84af93f

Commits on Jul 11, 2023

• fix attention_mask for chatglm (#6362)

  wj-Mcat

  authored
  6f1c0bc

  Copy full SHA for 6f1c0bc

• [llm]update llama & bloom &chatglm example (#6329)

  Show description for c33b59e
  lugimzzz

  authored
  c33b59e

  Copy full SHA for c33b59e

• [llama]fix tensor parallel output (#6344)

  Show description for c2c30e3
  lugimzzz

  authored
  c2c30e3

  Copy full SHA for c2c30e3

• Add bloom detail results (#6336)

  w5688414

  authored
  6267e9b

  Copy full SHA for 6267e9b

Commits on Jul 10, 2023

• [NewFeature] shift-label features (#6328)

  Show description for 3c883df
  wj-Mcat

  authored
  3c883df

  Copy full SHA for 3c883df