Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning
Paper
•
2503.09516
•
Published
•
36
Search-R1
Preliminary checkpoints with outcome-only RL.
-
-
PeterJinGo/SearchR1-nq_hotpotqa_train-llama3.2-3b-em-ppo
4B • Updated • 7 -
PeterJinGo/SearchR1-nq_hotpotqa_train-llama3.2-3b-em-grpo
4B • Updated • 4 -
PeterJinGo/SearchR1-nq_hotpotqa_train-llama3.2-3b-it-em-ppo
4B • Updated • 2 -
PeterJinGo/SearchR1-nq_hotpotqa_train-llama3.2-3b-it-em-grpo
4B • Updated • 8 -
PeterJinGo/SearchR1-nq_hotpotqa_train-qwen2.5-3b-em-ppo
3B • Updated • 126 -
PeterJinGo/SearchR1-nq_hotpotqa_train-qwen2.5-3b-em-grpo
3B • Updated • 91 -
PeterJinGo/SearchR1-nq_hotpotqa_train-qwen2.5-3b-it-em-ppo
3B • Updated • 22 • 1 -
PeterJinGo/SearchR1-nq_hotpotqa_train-qwen2.5-3b-it-em-grpo
3B • Updated • 36 -
PeterJinGo/SearchR1-nq_hotpotqa_train-qwen2.5-7b-em-ppo
8B • Updated • 2.58k • 3 -
PeterJinGo/SearchR1-nq_hotpotqa_train-qwen2.5-7b-it-em-ppo
8B • Updated • 60 -
PeterJinGo/wiki-18-corpus
Updated • 3.51k -
PeterJinGo/wiki-18-e5-index
Updated • 1.69k -
PeterJinGo/nq_hotpotqa_train
Viewer • Updated • 221k • 1.24k • 8 -
PeterJinGo/LICENCE
Viewer • Updated • 202 • 23