Examples - a trl-lib Collection

trl-lib 's Collections

Papers implemented in TRL

Prompt datasets (for GRPO, RLOO, ...)

Preference datasets (for DPO, reward modeling)

Unpaired preference datasets (KTO)

Models from the docs

Comparing DPO with IPO and KTO

Examples

updated 18 minutes ago

Base models, data and trained models from the TRL examples: https://github.com/huggingface/trl/tree/main/examples

LiquidAI/LFM2.5-8B-A1B

Text Generation • 8B • Updated 9 days ago • 21.7k • 809

Note LFM2.5 research agent: base model
argilla/apigen-function-calling

Viewer • Updated Oct 10, 2024 • 109k • 9.41k • 22

Note LFM2.5 research agent: SFT data
trl-lib/LFM2.5-8B-A1B-SFT-function-calling

Text Generation • 8B • Updated 28 days ago • 277

Note LFM2.5 research agent: after SFT
hotpotqa/hotpot_qa

Viewer • Updated Aug 11, 2025 • 203k • 87.6k • 363

Note LFM2.5 research agent: GRPO data
trl-lib/LFM2.5-8B-A1B-GRPO-HotpotQA

Text Generation • 8B • Updated 28 days ago • 289

Note LFM2.5 research agent: after GRPO
openai/gpt-oss-20b

Text Generation • 21B • Updated Aug 26, 2025 • 5.69M • • 5.3k

Note gpt-oss chess: base model
trl-lib/chess-puzzles

Viewer • Updated 9 days ago • 6.16M • 128

Note gpt-oss chess: data, built from Lichess/chess-puzzles
trl-lib/gpt-oss-20b-chess

Text Generation • 21B • Updated 9 days ago • 246 • 1

Note gpt-oss chess: after SFT