RAGLAB: A Modular and Research-Oriented Unified Framework for Retrieval-Augmented Generation

• RAGLAB is a modular, research-oriented open-source framework for Retrieval-Augmented Generation (RAG) algorithms. It offers reproductions of 6 existing RAG algorithms and a comprehensive evaluation system with 10 benchmark datasets, enabling fair comparisons between RAG algorithms and easy expansion for efficient development of new algorithms, datasets, and evaluation metrics.

🌟Features

• Comprehensive RAG Ecosystem: Supports the entire RAG pipeline from data collection and training to auto-evaluation.
• Advanced Algorithm Implementations: Reproduces 6 state-of-the-art RAG algorithms, with an easy-to-extend framework for developing new algorithms.
• Fair Comparison Platform: Provides benchmark results for 6 algorithms across 5 task types and 10 datasets.
• Efficient Retriever Client: Offers local API for parallel access and caching, with average latency under 1 second.
• Versatile Generator Support: Compatible with 70B+ models, VLLM, and quantization techniques.
• Flexible Instruction Lab: Customizable instruction templates for various RAG scenarios.

🔨Install environment

• dev environment：pytorch:2.0.1-py3.10-cuda11.8.0-devel-ubuntu22.04

• install miniconda

• git clone raglab

  git clone https://github.com/fate-ubw/raglab-exp.git

• create environment from yml file

  cd raglab-exp
  conda env create -f environment.yml

• install flash-attn, en_core_web_sm, punkt manually

  pip install flash-attn==2.2
  python -m spacy download en_core_web_sm
  python -m nltk.downloader punkt

🤗 Model

• raglab need llama2-7b, llama3-8b, colbertv2.0, selfrag_llama2_7b

  cd raglab-exp
  mkdir model
  cd model
  mkdir output_models
  mkdir Llama-2-7b-hf
  huggingface-cli download meta-llama/Llama-2-7b-hf --local-dir Llama-2-7b-hf/
  mkdir Meta-Llama-3-8B
  huggingface-cli download meta-llama/Meta-Llama-3-8B --local-dir Meta-Llama-3-8B/
  mkdir Meta-Llama-3-70B
  huggingface-cli download meta-llama/Meta-Llama-3-70B --local-dir Meta-Llama-3-70B/
  mkdir selfrag_llama2_7b
  huggingface-cli download selfrag/selfrag_llama2_7b --local-dir selfrag_llama2_7b/
  mkdir colbertv2.0
  huggingface-cli download colbert-ir/colbertv2.0 --local-dir colbertv2.0/

💽 process wiki2023 as vector database

10-samples test

• 10-samples test is aimed at validating the environment
• run colbert embedding process enwiki-20230401-10samples.tsv
  1. Change root path for variables: checkpoint, index_dbPath, collection in wiki2023-10samples_tsv-2-colbert_embedding.py. In file paths, colbert encounters many issues when using relative paths to generate embeddings. Therefore, the current version of raglab uses absolute paths.

    # change root path
  checkpoint = '/your_root_path/raglab-exp/model/colbertv2.0'
  index_dbPath = '/your_root_path/raglab-exp/data/retrieval/colbertv2.0_embedding/wiki2023-10samples'
  collection = '/your_root_path/raglab-exp/data/retrieval/colbertv2.0_passages/wiki2023-10samples/enwiki-20230401-10samples.tsv'

  2. run

  cd raglab-exp
  sh run/wiki2023_preprocess/2-wiki2023-10samples_tsv-2-colbert_embedding.sh

• Embedding precess will take around 15mins in first time.
• The first time colbert processes embeddings, it takes a relatively long time because it needs to recompile the torch_extensions. However, calling the processed embeddings does not require a long time. If there are no errors and the retrieved text can be printed, it indicates that the environment is correct.

Run Raglab with 10-samples embedding

• test selfrag base on 10-samples embedding
• After processing with colbert embeddings, you can start running the algorithms in raglab. All algorithms integrated in raglab include two modes: interact and evaluation. The test stage demonstrates in interact mode, just for fun 🤗.
• Modify the index_dbPath and text_dbPath in config file:selfrag_reproduction-interact-short_form-adaptive_retrieval.yaml

  index_dbPath: /your_root_path/raglab-exp/data/retrieval/colbertv2.0_embedding/wiki2023-10samples
  text_dbPath: /your_root_path/raglab-exp/data/retrieval/colbertv2.0_passages/wiki2023-10samples/enwiki-20230401-10samples.tsv

• run selfrag (short form & adaptive retrieval) interact mode test 10-samples embedding

  cd raglab-exp
  sh run/rag_inference/3-selfrag_reproduction-interact-short_form-adaptive_retrieval.sh

• Congratulations！！！Now you have already know how to run raglab 🌈
• In raglab, each algorithm has 10 queries built-in in interact mode which are sampled from benchmark

embedding whole wiki2023

• you can download the colbert embdding wiki2023 as raglab database(40Gb)

cd /raglab-exp/data/retrieval/colbertv2.0_embedding
gdown --id xxxxxx
# unzip commend for 

• modify the path in meta.json file
• embedding whole wiki2023 to vector need 22 hours, so we recommend download prepared embedding

download wiki2023 raw data

• current version of raglab use wiki2023 as database
• we get source wiki2023 get from factscore
  • method1: url for download wiki2023:google_drive
  • method2: install throuth gdown

  cd raglab-exp/data/retrieval/colbertv2.0_passages
  mkdir wiki2023
  pip install gdown
  gdown --id 1mekls6OGOKLmt7gYtHs0WGf5oTamTNat

preprocess wiki2023

• If the 10-samples test is passed successfully, you can proceed with processing wiki2023.

1. preprocess .db -> .tsv (Colbert can only read files in .tsv format.)

   cd raglab-exp
   sh run/wiki2023_preprocess/3-wiki2023_db-2-tsv.sh

2. .tsv -> embedding

• remember to change the root path of checkpoint, index_dbPath and collection

    # change root path
      checkpoint = '/your_root_path/raglab-exp/model/colbertv2.0'
      index_dbPath = '/your_root_path/raglab-exp/data/retrieval/colbertv2.0_embedding/wiki2023-10samples'
      collection = '/your_root_path/raglab-exp/data/retrieval/colbertv2.0_passages/wiki2023-10samples/enwiki-20230401-10samples.tsv'

• run bash script

  cd raglab-exp
  sh run/wiki2023_preprocess/4-wiki2023_tsv-2-colbert_embedding.sh

💽 Process wiki2018 as vector database

• This section is a tutorial on using wiki2018

Download text files

• Directly download wiki2018 raw database using wget

cd raglab-exp/data/retrieval/colbertv2.0_passages/wiki2018
wget https://dl.fbaipublicfiles.com/dpr/wikipedia_split/psgs_w100.tsv.gz

Process raw wiki2018 into colbert format

cd raglab-exp
sh run/wiki2018_preprocess/1-wiki2018_tsv_2_tsv.sh

Modify wiki2018 embedding config file

1. Change the path

cd /raglab-exp/data/retrieval/colbertv2.0_embedding/wiki2018/indexes/wiki2018
vim metadata.json 

• You only need to modify two paths in the metadata.json file. Here, simply delete the original paths and copy the following paths. Other parameters do not need to be modified.

"collection": "/home/ec2-user/SageMaker/raglab-exp/data/retrieval/colbertv2.0_passages/wiki2018/wiki2018.tsv",
"experiment": "/home/ec2-user/SageMaker/raglab-exp/data/retrieval/colbertv2.0_embedding/wiki2018",

• After modification, you can directly start the colbert server. For experimental startup method, refer to the last section of the readme: Inference experiments.

Inference experiments

Retrieval server & api

• The inference experiments require running hundreds of scripts in parallel. If each script loads the wiki2023 database separately, not only does it require a large amount of RAM, but loading the wiki2023 database each time also takes a considerable amount of time, which is a significant waste of computing resources. Therefore, RagLab has designed colbert server & colbert api to address the problem of multi-task parallel retrieval. By runnging local colbert server, tasks can call the colbert api to obtain retrieval results, greatly reducing the inference time for multiple tasks.
• Attention: colbert_server need atleast 60GB ram

  cd raglab-exp
  sh run/colbert_server/colbert_server.sh

• open another terminal test your ColBERT server

cd raglab-exp
sh run/colbert_server/ask_api.sh

• ColBERT server started successfully!!! 🌈

Automatic GPU Scheduler

• inference experiments require running hundreds of scripts in parallel, the automatic gpu scheduler needs to be used to automatically allocate GPUs for different bash scripts in Parallel.
• install simple_gpu_scheduler

  pip install simple_gpu_scheduler

• run hundreds of experiments in one line 😎

  cd raglab-exp
  simple_gpu_scheduler --gpus 0,1,2,3,4,5,6,7 < auto_gpu_scheduling_scripts/auto_run_scripts-jeff.py

• how to write your_script.txt?
  • here is an example

  # auto_inference_selfreg-7b.txt
  sh run/rag_inference/selfrag_reproduction/selfrag_reproduction-evaluation-short_form-PubHealth-adaptive_retrieval-pregiven_passages.sh
  sh run/rag_inference/selfrag_reproduction/selfrag_reproduction-evaluation-short_form-PubHealth-always_retrieval-pregiven_passages.sh

Fine tune llama3 & self rag

• The base models for raglab baseline and selfrag use llama3-instruction-8b. Since selfrag was further fine-tuned on additional data during the fine-tuning stage, in order to make a fair comparison, the baseline model also needs to be fine-tuned.

download self rag train data

• we get the train data from selfrag
• google drive url
• download through gdown

  cd raglab-exp/data/train_data/
  gdown --id 10G_FozUV4u27EX0NjwVe-3YMUMeTwuLk

10-samples test for fintune

• The 10-samples train dataset has been processed, please directly start the bash script to begin testing.
• Note: The test script only uses one GPU
  • full weight requires 80GB VRam GPU

  cd raglab-exp
  sh run/rag_train/script_finetune-llama3-baseline-full_weight-10samples.sh

  • LoRA (Low-Rank Adaptation) requires at least 26GB of VRAM

  cd raglab-exp
  sh run/rag_train/script_finetune-llama3-baseline-Lora-10samples.sh

• Congratulations！！！You can now start fine-tuning the baseline and selfrag-8b🤖

finetune self rag 8b

• full weight finetune

  cd raglab-exp
  sh run/rag_train/script_finetune-selfrag_8b-full_weight.sh

• lora finetune

  cd raglab-exp
  sh run/rag_train/script_finetune-selfrag_8b-Lora.sh

finetune llama3-8b as baseline

• preprocess train data. Train data for baseline model need remove special tokens.

  cd raglab-exp
  sh run/traindataset_preprocess/selfrag_traindata-remove_special_tokens.sh

• then you will get baseline train_data without special token and passages (Q: what is specal token? Anawer: special tokens is a concept proposed by SelfRAG)
• full weight finetune llama3-8b-baseline ues processed data

  sh run/rag_train/script_finetune-llama3-baseline-full_weight.sh

• lora finetune llama3-8b-baseline

  cd raglab-exp
  sh run/rag_train/script_finetune-llama3-baseline-Lora.sh

Lora finetune llama3-70b as baseline

• preprocess train data. Train data for baseline model need remove special tokens.

  cd raglab-exp
  sh run/traindataset_preprocess/selfrag_traindata-remove_special_tokens.sh

• lora finetune llama3-70b-baseline ues processed data

  sh run/rag_train/script_finetune-llama3-70B-baseline-Lora.sh

QLora finetune llama3-70B as baseline

• preprocess train data. Train data for baseline model need remove special tokens.

  cd raglab-exp
  sh run/traindataset_preprocess/selfrag_traindata-remove_special_tokens.sh

• 8bit QLora finetune llama3-70B

  sh run/rag_train/script_finetune-llama3-70B-baseline-QLora-8bit.sh

• 4bit QLora fintune llama3-70B

  sh run/rag_train/script_finetune-llama3-70B-baseline-QLora-4bit.sh

8bit QLora finetune selfrag-70B as baseline

• 8bit Qlora finetune slefrag 70B

    sh run/rag_train/script_finetune-selfrag_llama3-70b-QLora-8bit.sh

• 4bit Qlora finetune slefrag 70B

    sh run/rag_train/script_finetune-selfrag_llama3-70b-QLora-4bit.sh

🔖 License

FlashRAG is licensed under the MIT License.