性能概览：

什么是 s3？

s3 框架

s3 是一个用于训练检索增强生成（RAG）搜索代理的简单而强大的框架。它教会语言模型如何更有效地进行搜索——无需更改生成器本身。通过仅关注搜索组件，s3 用远少于以往方法的数据实现了在 QA 任务上的强劲性能。它具有模块化、高效的特点，并且可以与任何黑盒 LLM 无缝协作。

目录

• 📦 安装
• 💡 准备
• 🏋️ 开始训练
• 🔍 执行搜索/检索
• 📈 运行评估

📦 安装

搜索器与生成器环境

conda create -n s3 python=3.9
install torch [or you can skip this step and let vllm to install the correct version for you]
pip install torch==2.4.0 --index-url https://download.pytorch.org/whl/cu121
install vllm
pip3 install vllm==0.6.3 # or you can install 0.5.4, 0.4.2 and 0.3.1
pip3 install ray
verl
cd code
pip install -e .
flash attention 2
pip3 install flash-attn --no-build-isolation
we use pyserini for efficient retrieval and evaluation
pip install pyserini    # the version we used is 0.22.1
quality of life
pip install wandb IPython matplotlib huggingface_hub

conda create -n ret python=3.10
conda activate ret
conda install pytorch==2.4.0 torchvision==0.19.0 torchaudio==2.4.0 pytorch-cuda=12.1 -c pytorch -c nvidia
pip install transformers datasets pyserini
conda install -c pytorch -c nvidia faiss-gpu=1.8.0
pip install uvicorn fastapi

💡 准备工作

python scripts/download.py --save_path $save_path
cat $save_path/part_* > $save_path/e5_Flat.index
gzip -d $save_path/wiki-18.jsonl.gz

预计算 Naïve RAG 初始化（或者你可以在这里下载我们处理过的数据：huggingface）

# deploy retriever
bash scripts/deploy_retriever/retrieval_launch.sh # or scripts/deploy_retriever/retrieval_launch_mirage.sh for MedCorp corpus.
deploy generator
bash generator_llms/host.sh # modify tensor-parallel-size to the number of GPUs you use
run precompute
bash scripts/precompute.sh # this step will take a while, as it will precompute the naïve RAG Cache for training

🏋️ 运行训练

# deploy retriever
bash scripts/deploy_retriever/retrieval_launch.sh 
deploy generator
bash generator_llms/host.sh
run training
bash scripts/train/train_s3.sh

🔍 运行搜索/检索

s3

# deploy retriever
bash scripts/deploy_retriever/retrieval_launch.sh 
run s3 inference
bash scripts/s3_inference/evaluate-8-3-3.sh

bash scripts/deploy_retriever/retrieval_launch.sh # or retrieval_launch_bm25.sh # deploy retriever
bash scripts/baselines/rag.sh # run RAG 

bash retrieval_launch_bm25.sh # deploy BM25 Model
bash generator_llms/deepretrieval.sh # deploy DeepRetrieval Model
bash scripts/baselines/deepretrieval.sh # run DeepRetrieval Query Rewriting + Retrieval

bash retrieval_launch.sh # deploy e5 retriever
bash scripts/baselines/search_r1.sh # run Search-R1

bash retrieval_launch.sh # deploy e5 retriever
python scripts/baselines/ircot.py

bash retrieval_launch.sh # deploy e5 retriever
bash scripts/baselines/search_o1.sh # run Search-o1

📈 运行评估

bash scripts/evaluation/run.sh

问答

定制数据？

复现结果？

虽然自行复现模型其实很简单——我们实际上推荐从零开始训练，因为评估往往比训练更耗时——但我们也提供了一个参考模型检查点：s3-8-3-3-20steps，训练时间约为一小时。

鸣谢

引用

@article{jiang2025s3,
  title={s3: You Don't Need That Much Data to Train a Search Agent via RL},
  author={Jiang, Pengcheng and Xu, Xueqiang and Lin, Jiacheng and Xiao, Jinfeng and Wang, Zifeng and Sun, Jimeng and Han, Jiawei},
  journal={arXiv preprint arXiv:2505.14146},
  year={2025}
}

--- Tranlated By Open Ai Tx | Last indexed: 2025-10-06 ---