Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Jul 27, 2023

• Add missing max_seq_length arg to example sft_trainer.py (huggingface#585)

  SharkWipf

  authored
  3b0a1b5

• Computes the KL penalty using the entire distribution (huggingface#541)

  Show description for 31658b4
  edbeeching

  authored
  31658b4

Commits on Jul 26, 2023

• Fix model output dim in reward trainer example (huggingface#566)

  Show description for f7227fb
  liutianlin0121

  authored
  f7227fb

• [DPO] Resolve logging for DPOTrainer (huggingface#570)

  Show description for b3c2e73
  tomaarsen

  authored
  b3c2e73

Commits on Jul 24, 2023

• Add comment to explain how the sentiment pipeline is used to run the … (huggingface#555)

  Show description for d78d917
  jvhoffbauer

  authored
  d78d917

• Add DataCollatorForCompletionOnlyLM in the docs (huggingface#565)

  Show description for cdde7f7
  younesbelkada

  authored
  cdde7f7

• add epochs and num steps on CLI (huggingface#562)

  younesbelkada

  authored
  51d5f08

• Minor typo and whitespace fixes (huggingface#559)

  Show description for 8762507
  tmm1

  authored
  8762507

• remove unused batch_size arg (huggingface#554)

  kashif

  authored
  1bd852a

Commits on Jul 22, 2023

• [SFTTrainer] Add warning for wrong padding_side (huggingface#550)

  Show description for 170d58f
  younesbelkada

  authored
  170d58f

Commits on Jul 20, 2023

• ADD: num_proc to SFTTrainer (huggingface#547)

  Show description for 84c9209
  

  3 peopleauthored
  84c9209

Commits on Jul 19, 2023

• Add use_auth_token arg to sft_trainer example (huggingface#544)

  Show description for d0fe348
  

  corey-lambda

  and

  lvwerra

  authored
  d0fe348

• [examples] make the sft script more modulable (huggingface#543)

  Show description for 5857d0a
  younesbelkada

  authored
  5857d0a

• [DPO] remove response/pairs from the DPO side (huggingface#540)

  Show description for fd50e06
  

  3 peopleauthored
  fd50e06

• Relax reward trainer constraint (huggingface#539)

  Show description for bcff7c2
  

  younesbelkada

  and

  tomaarsen

  authored
  bcff7c2

• fix offline case (huggingface#538)

  younesbelkada

  authored
  0e8d9f8

Commits on Jul 18, 2023

• all the concated batches are on same device (huggingface#528)

  kashif

  authored
  7f297b3

Commits on Jul 17, 2023

• DPO Trainer (huggingface#416)

  Show description for 84393f3
  

  4 peopleauthored
  84393f3

Commits on Jul 14, 2023

• Fix sentiment nit (huggingface#517)

  younesbelkada

  authored
  388bdc0

• [examples] Big refactor of examples and documentation (huggingface#509)

  Show description for 5c7bfbc
  younesbelkada

  authored
  5c7bfbc

• Use local process index for _get_current_device() (huggingface#515)

  Show description for 36b77ae
  lewtun

  authored
  36b77ae

Commits on Jul 13, 2023

• Put labels tensors onto GPU to fix eval bug on deepspeed (huggingface#513)

  rlindskog

  authored
  2049d03

• set dev version

  younesbelkada

  committed
  31b98aa

• git commit -m 'Release: v0.4.7'

  younesbelkada

  committed
  d06b131

Commits on Jul 12, 2023

• [SFTTrainer] Fix the sequence length check of SFTTrainer (huggingface#512)

  Show description for f323090
  

  3 peopleauthored
  f323090

Commits on Jul 6, 2023

• [PPOTrainer] Add prompt tuning support on TRL (huggingface#500)

  Show description for bbc7eeb
  younesbelkada

  authored
  bbc7eeb

• [PPOTrainer] Add prefix tuning support (huggingface#501)

  Show description for 163dae5
  younesbelkada

  authored
  163dae5

• Update ppo_trainer.py (huggingface#499)

  younesbelkada

  authored
  64c8db2

• Disable mlm by default in DataCollatorForCompletionOnlyLM, add ignore_index and docstring (huggingface#476)

  Show description for 25d4d81
  

  BramVanroy

  and

  Bram Vanroy

  authored
  25d4d81

• correctly implement gradient checkpointing (huggingface#479)

  Show description for 685620a
  mnoukhov

  authored
  685620a

Commits on Jul 5, 2023

• Adds some options to stabilize the KL penalty (huggingface#486)

  Show description for 2b531b9
  

  3 peopleauthored
  2b531b9

• Remove padding in batched generation. (huggingface#487)

  Show description for 4f7f73d
  

  lvwerra

  and

  leandro von werra

  authored
  4f7f73d

Commits on Jul 4, 2023

• FIX: contributing guidelines command (huggingface#493)

  Show description for c60c416
  

  BramVanroy

  and

  lvwerra

  authored
  c60c416

• fix typo in reward_modeling.py (huggingface#494)

  csyourui

  authored
  cbb98da

• add ratio threshold to avoid spikes (huggingface#488)

  lvwerra

  authored
  a86eaab

Pagination

PreviousNext