Commits

Branch selector

main

User selector

All users

All time

Commit history

Commits on Oct 8, 2024

• Get the `aux_loss_coef` at `BCOTrainer`, `CPOTrainer`, `KTOTrainer`, and `ORPOTrainer` initialization (#2201)

  Show description for 511c92c
  muupan

  authored
  511c92cView commit details

  Copy full SHA for 511c92c

  Browse repository at this point

• Get the `aux_loss_coef` at `DPOTrainer` initialization (#2200)

  muupan

  authored
  c6cb635View commit details

  Copy full SHA for c6cb635

  Browse repository at this point

• ♾️ [CI] Use transformers from source in "tests_no_optional_dep" (#2198)

  qgallouedec

  authored
  adb3e05View commit details

  Copy full SHA for adb3e05

  Browse repository at this point

Commits on Oct 7, 2024

• `skip_prompt=True` in `TextIteratorStreamer` (#2193)

  Show description for adf58d8
  

  qgallouedec

  and

  kashif

  authored
  adf58d8View commit details

  Copy full SHA for adf58d8

  Browse repository at this point

• Update README.md (#2186)

  Show description for 9aa0225
  

  3 peopleauthored
  9aa0225View commit details

  Copy full SHA for 9aa0225

  Browse repository at this point

• Fix RLOO checkpointing (#2114)

  Show description for 82ad390
  

  3 peopleauthored
  82ad390View commit details

  Copy full SHA for 82ad390

  Browse repository at this point

• Update CONTRIBUTING.md (#2181)

  Show description for ac038ef
  

  kushal34712

  and

  kashif

  authored
  ac038efView commit details

  Copy full SHA for ac038ef

  Browse repository at this point

• [CI] fix dpo gpu ci tests (#2189)

  Show description for 51ca76b
  kashif

  authored
  51ca76bView commit details

  Copy full SHA for 51ca76b

  Browse repository at this point

• 🃏 Model card: "unsloth" tag (#2173)

  qgallouedec

  authored
  7005ab4View commit details

  Copy full SHA for 7005ab4

  Browse repository at this point

• Update documentation CLI Chat (#2191)

  qgallouedec

  authored
  ffb1ab7View commit details

  Copy full SHA for ffb1ab7

  Browse repository at this point

• Rename trainer arg `tokenizer` to `processing_class` (#2162)

  qgallouedec

  authored
  47d08a9View commit details

  Copy full SHA for 47d08a9

  Browse repository at this point

• add trl to tag for models (#2178)

  qgallouedec

  authored
  70327c1View commit details

  Copy full SHA for 70327c1

  Browse repository at this point

Commits on Oct 6, 2024

• minor KTO setting changes + KL batch size (#2153)

  Show description for f05c3fa
  

  kawine

  and

  kashif

  authored
  f05c3faView commit details

  Copy full SHA for f05c3fa

  Browse repository at this point

Commits on Oct 5, 2024

• Capybara replaced with ultrafeedback_binarized (#2183)

  August-murr

  authored
  4799ba4View commit details

  Copy full SHA for 4799ba4

  Browse repository at this point

Commits on Oct 4, 2024

• Conversational dataset support for `CPOTrainer` (#2144)

  Show description for d45c86e
  

  qgallouedec

  and

  lewtun

  authored
  d45c86eView commit details

  Copy full SHA for d45c86e

  Browse repository at this point

• 🗑️ Set deprecation version for DPO and SFT arguments to version 0.13 (#2170)

  qgallouedec

  authored
  c6b0d13View commit details

  Copy full SHA for c6b0d13

  Browse repository at this point

• Update trl version in CITATION.cff (#2171)

  qgallouedec

  authored
  3321084View commit details

  Copy full SHA for 3321084

  Browse repository at this point

• Default `dataset_text_field` to `"text"` (#2078)

  Show description for a9cffc7
  qgallouedec

  authored
  a9cffc7View commit details

  Copy full SHA for a9cffc7

  Browse repository at this point

• 🏷️ Model badges in trainer documentation (#2160)

  qgallouedec

  authored
  32a928cView commit details

  Copy full SHA for 32a928c

  Browse repository at this point

• Fix typo in error message (#2168)

  Show description for 1a3bb37
  eltociear

  authored
  1a3bb37View commit details

  Copy full SHA for 1a3bb37

  Browse repository at this point

Commits on Oct 3, 2024

• ↩️ Revert tokenizer hotfix #2163

  qgallouedec

  authored
  d4564b7View commit details

  Copy full SHA for d4564b7

  Browse repository at this point

• 🩹 [Hotfix] Add setter for tokenizer (#2163)

  qgallouedec

  authored
  1be4d86View commit details

  Copy full SHA for 1be4d86

  Browse repository at this point

Commits on Oct 2, 2024

• Conversational dataset support for `DPOTrainer` (#2131)

  Show description for 78249d9
  

  3 peopleauthored
  78249d9View commit details

  Copy full SHA for 78249d9

  Browse repository at this point

Commits on Oct 1, 2024

• [CI] Don't use `eval_strategy="steps"` when no eval dataset (#2152)

  Show description for 5c21de3
  qgallouedec

  authored
  5c21de3View commit details

  Copy full SHA for 5c21de3

  Browse repository at this point

• 🩹 Fix attention mask warning in chat CLI (#2147)

  Show description for 0a566f0
  qgallouedec

  authored
  0a566f0View commit details

  Copy full SHA for 0a566f0

  Browse repository at this point

Commits on Sep 30, 2024

• [GKD] Set custom EOS tokens in generation config (#2142)

  Show description for de38765
  

  lewtun

  and

  qgallouedec

  authored
  de38765View commit details

  Copy full SHA for de38765

  Browse repository at this point

Commits on Sep 27, 2024

• rename example (#2139)

  qgallouedec

  authored
  1201aa6View commit details

  Copy full SHA for 1201aa6

  Browse repository at this point

• 🃏 Model card for TRL (#2123)

  Show description for c00722c
  

  qgallouedec

  and

  lewtun

  authored
  c00722cView commit details

  Copy full SHA for c00722c

  Browse repository at this point

• Add correct label for `WinRateCallback` table (#2134)

  Show description for 124189c
  lewtun

  authored
  124189cView commit details

  Copy full SHA for 124189c

  Browse repository at this point

• arXiv to HF papers (#2133)

  qgallouedec

  authored
  d5eeaabView commit details

  Copy full SHA for d5eeaab

  Browse repository at this point

Commits on Sep 26, 2024

• 🧹 Style (#2132)

  Show description for 5368be1
  qgallouedec

  authored
  5368be1View commit details

  Copy full SHA for 5368be1

  Browse repository at this point

• Add table for WinRateCallback (#2116)

  Show description for b169e10
  

  lewtun

  and

  qgallouedec

  authored
  b169e10View commit details

  Copy full SHA for b169e10

  Browse repository at this point

• ♻️ Standardize `script_args` (#2130)

  qgallouedec

  authored
  9af4734View commit details

  Copy full SHA for 9af4734

  Browse repository at this point

• Tokenize row during in `training_step` in `OnlineDPOTrainer` (#2117)

  Show description for a0d7149
  

  qgallouedec

  and

  lewtun

  authored
  a0d7149View commit details

  Copy full SHA for a0d7149

  Browse repository at this point

• Eos token encouragement Clarification (#2128)

  Show description for a0e2814
  August-murr

  authored
  a0e2814View commit details

  Copy full SHA for a0e2814

  Browse repository at this point