Commits

Branch selector

main

User selector

All users

All time

Commit history

Commits on Nov 29, 2024

• ⚠️ Add warning guidelines and update codebase to follow best practices (#2350)

  Show description for d6a8f2c
  qgallouedec

  authored
  d6a8f2cView commit details

  Copy full SHA for d6a8f2c

  Browse repository at this point

• 🌋 Add support for LLaVA-Next in `DPOTrainer` (#2413)

  Show description for 8d9cfaa
  

  4 peopleauthored
  8d9cfaaView commit details

  Copy full SHA for 8d9cfaa

  Browse repository at this point

• 🔓 Remove lm_head check in `AutoModelForCausalLMWithValueHead` (#2398)

  Show description for 94e4135
  qgallouedec

  authored
  94e4135View commit details

  Copy full SHA for 94e4135

  Browse repository at this point

• 🌐 Community Tutorials (#2411)

  Show description for ac26778
  

  burtenshaw

  and

  qgallouedec

  authored
  ac26778View commit details

  Copy full SHA for ac26778

  Browse repository at this point

• Add note about special tokens in chat templates for LoRA SFT (#2414)

  lewtun

  authored
  2c6e0d9View commit details

  Copy full SHA for 2c6e0d9

  Browse repository at this point

Commits on Nov 28, 2024

• 👁️ Added SFT support for `SmolVLM` models via standalone script `sft_vlm_smol_vlm.py` (#2409)

  Show description for e1d7813
  sergiopaniego

  authored
  e1d7813View commit details

  Copy full SHA for e1d7813

  Browse repository at this point

• 🖨 Add Script Utilities section to the documentation (#2407)

  Show description for a34e9bf
  qgallouedec

  authored
  a34e9bfView commit details

  Copy full SHA for a34e9bf

  Browse repository at this point

Commits on Nov 26, 2024

• 🗝️ Update type hints (#2399)

  Show description for c10cc89
  qgallouedec

  authored
  c10cc89View commit details

  Copy full SHA for c10cc89

  Browse repository at this point

• 🐢 Fix slow tests (#2397)

  Show description for 9368dcc
  

  kashif

  and

  qgallouedec

  authored
  9368dccView commit details

  Copy full SHA for 9368dcc

  Browse repository at this point

• 🧳 Move zen generation script and fix tests (#2393)

  Show description for 43df3a4
  

  qgallouedec

  and

  kashif

  authored
  43df3a4View commit details

  Copy full SHA for 43df3a4

  Browse repository at this point

• 🖋️ Fix warning message formatting in `KTOTrainer` (#2394)

  qgallouedec

  authored
  baee06fView commit details

  Copy full SHA for baee06f

  Browse repository at this point

• 🤐 Fix deprecation warnings (#2395)

  qgallouedec

  authored
  bbd8cbbView commit details

  Copy full SHA for bbd8cbb

  Browse repository at this point

• 🤐 Fix deprecation warnings (#2392)

  qgallouedec

  authored
  4f937c7View commit details

  Copy full SHA for 4f937c7

  Browse repository at this point

• 👮 Deprecate `policy` in favor of `model` in `PPOTrainer` (#2386)

  qgallouedec

  authored
  16fa13cView commit details

  Copy full SHA for 16fa13c

  Browse repository at this point

Commits on Nov 25, 2024

• 🤏 New models for tests (#2287)

  Show description for 453db5c
  qgallouedec

  authored
  453db5cView commit details

  Copy full SHA for 453db5c

  Browse repository at this point

• 💾 Deprecate `config` in favor of `args` in `PPOTrainer` (#2384)

  qgallouedec

  authored
  ee3cbe1View commit details

  Copy full SHA for ee3cbe1

  Browse repository at this point

• 📦 Support for packing tokenized datasets for SFT (#2011)

  Show description for 17e8060
  kmehant

  authored
  17e8060View commit details

  Copy full SHA for 17e8060

  Browse repository at this point

Commits on Nov 24, 2024

• 🙈 Suppress warning for estimating tokens in trainers (#2389)

  Show description for 163695e
  qgallouedec

  authored
  163695eView commit details

  Copy full SHA for 163695e

  Browse repository at this point

Commits on Nov 21, 2024

• Update log method to include start_time parameter (#2381)

  qgallouedec

  authored
  672c965View commit details

  Copy full SHA for 672c965

  Browse repository at this point

• 📝 Fix typo in dataset generation script (#2379)

  jiseshen

  authored
  bdeb117View commit details

  Copy full SHA for bdeb117

  Browse repository at this point

• 🔀 Add `MergeModelCallBack` (#2282)

  Show description for 6578fdc
  

  4 peopleauthored
  6578fdcView commit details

  Copy full SHA for 6578fdc

  Browse repository at this point

Commits on Nov 20, 2024

• ⏰ Add `start_time` to `_maybe_log_save_evaluate` (#2373)

  qgallouedec

  authored
  a0066f4View commit details

  Copy full SHA for a0066f4

  Browse repository at this point

• 🧲 Use our own `require_bitsandbytes` (#2370)

  Show description for 5626806
  qgallouedec

  authored
  5626806View commit details

  Copy full SHA for 5626806

  Browse repository at this point

• remove redunant call to eval and train (#2372)

  kashif

  authored
  bb0afc2View commit details

  Copy full SHA for bb0afc2

  Browse repository at this point

Commits on Nov 19, 2024

• Fix dev install (#2369)

  lewtun

  authored
  066fc37View commit details

  Copy full SHA for 066fc37

  Browse repository at this point

Commits on Nov 18, 2024

• 🎲 Move random judges in testing utilities (#2365)

  Show description for b80c1a6
  qgallouedec

  authored
  b80c1a6View commit details

  Copy full SHA for b80c1a6

  Browse repository at this point

• 🤝 Mixture of judges (#2159)

  Show description for b5eabbe
  

  gaetanlop

  and

  qgallouedec

  authored
  b5eabbeView commit details

  Copy full SHA for b5eabbe

  Browse repository at this point

• 🗺️ Implementation DiscoPOP Loss (#2323)

  Show description for cbf9abc
  

  fanconic

  and

  qgallouedec

  authored
  cbf9abcView commit details

  Copy full SHA for cbf9abc

  Browse repository at this point

• 📃 Fix description for parameter "generate_during_eval" in dpo_config (#2364)

  dakru012

  authored
  6f8fe59View commit details

  Copy full SHA for 6f8fe59

  Browse repository at this point

• 📉 Add PEFT support for `PPOTrainer` (#2344)

  Show description for 1293f37
  

  ccs96307

  and

  qgallouedec

  authored
  1293f37View commit details

  Copy full SHA for 1293f37

  Browse repository at this point

• 🗃️ Use specified `data_collator` in `RLOOTrainer` and `PPOTrainer` (#2360)

  Show description for e7870dd
  bartoszzuk

  authored
  e7870ddView commit details

  Copy full SHA for e7870dd

  Browse repository at this point

• 🔮 Inference mode in `GeometricMixtureWrapper.forward` (#2345)

  Show description for 21d5baf
  kashif

  authored
  21d5bafView commit details

  Copy full SHA for 21d5baf

  Browse repository at this point

• 🪜 Stepwise supervision dataset type (#2148)

  qgallouedec

  authored
  76dbb1aView commit details

  Copy full SHA for 76dbb1a

  Browse repository at this point

Commits on Nov 15, 2024

• ⚖️ Add `use_soft_judge` option to `WinRateCallback` (#2347)

  Show description for b8c9d9c
  

  kashif

  and

  qgallouedec

  authored
  b8c9d9cView commit details

  Copy full SHA for b8c9d9c

  Browse repository at this point

Commits on Nov 12, 2024

• 👋 Remove deprecated `tokenizer` argument in BCO, GKD, Iterative SFT, Nash MD and XPO (#2349)

  qgallouedec

  authored
  6239631View commit details

  Copy full SHA for 6239631

  Browse repository at this point