🧺 [2/N] Refactor `_generate` in GRPO/RLOO: Use `prompt_ids` from generation#4152

Merged

qgallouedec merged 92 commits intomainfrom

refactor_generate_2

Oct 7, 2025

Commits on Sep 20, 2025

Commits on Sep 25, 2025

Commits on Sep 26, 2025

Commits on Sep 30, 2025

Commits on Oct 6, 2025

🔣 Fix test: replace `trainer.tokenizer` by `trainer.processing_class` (#4185)
qgallouedec
committed
Fix CI ImportError: FlashAttention2 and decorator order for all parameterized tests (#4176)

albertvillanova
authored and
qgallouedec
committed
Hotfix wrong formatting of docstrings with blockquote tips (#4187)

albertvillanova
authored and
qgallouedec
committed
🌡️ Have vLLM return processed (temperature scaled) log probs (#4163)

YonatanGideoni
and
qgallouedec
committed
Replace remaining trainer.tokenizer with trainer.processing_class in GRPO test (#4192)

albertvillanova
authored and
qgallouedec
committed
[DOCS] Lora without regret (#4181)

committed
[DOCS/FIX] lora without regrets - fix lr (#4207)

burtenshaw
authored and
qgallouedec
committed
Remove custome_container for building the docs (#4198)

albertvillanova
authored and
qgallouedec
committed
Remove tokenizer creation from `sft` example script (#4197)

sergiopaniego
authored and
qgallouedec
committed
Hotfix: Exclude transformers 4.57.0 for Python 3.9 (#4209)

authored and
qgallouedec
committed
Replace unittest with pytest (#4188)

albertvillanova
authored and
qgallouedec
committed
Updated vLLM integration guide (#4162)

sergiopaniego
and
qgallouedec
committed
Remove `Optional` from `processing_class` in `PPOTrainer` (#4212)

sergiopaniego
authored and
qgallouedec
committed
Replace setup with pyproject and fix packaging unintended modules (#4194)

albertvillanova
authored and
qgallouedec
committed
Merge branch 'main' into refactor_generate_2
qgallouedec
authored
Merge branch 'main' into refactor_generate_2
qgallouedec
authored
Merge branch 'main' into refactor_generate_2
qgallouedec
authored

Commits on Oct 7, 2025