HippoRAG 2 is a powerful memory framework for LLMs that enhances their ability to recognize and utilize connections in new knowledgeβmirroring a key function of human long-term memory.
Our experiments show that HippoRAG 2 improves associativity (multi-hop retrieval) and sense-making (the process of integrating large and complex contexts) in even the most advanced RAG systems, without sacrificing their performance on simpler tasks.
Like its predecessor, HippoRAG 2 remains cost and latency efficient in online processes, while using significantly fewer resources for offline indexing compared to other graph-based solutions such as GraphRAG, RAPTOR, and LightRAG.
Figure 1: Evaluation of continual learning capabilities across three key dimensions: factual memory (NaturalQuestions, PopQA), sense-making (NarrativeQA), and associativity (MuSiQue, 2Wiki, HotpotQA, and LV-Eval). HippoRAG 2 surpasses other methods across all categories, bringing it one step closer to true long-term memory.
Figure 2: HippoRAG 2 methodology.
- HippoRAG: Neurobiologically Inspired Long-Term Memory for Large Language Models [NeurIPS '24].
- From RAG to Memory: Non-Parametric Continual Learning for Large Language Models [Under Review].
conda create -n hipporag python=3.10
conda activate hipporag
pip install hipporag
Initialize the environmental variables and activate the environment:
export CUDA_VISIBLE_DEVICES=0,1,2,3
export HF_HOME=<path to Huggingface home directory>
export OPENAI_API_KEY=<your openai api key> # if you want to use OpenAI model
conda activate hipporag
This simple example will illustrate how to use hipporag
with any OpenAI model:
from hipporag import HippoRAG
# Prepare datasets and evaluation
docs = [
"Oliver Badman is a politician.",
"George Rankin is a politician.",
"Thomas Marwick is a politician.",
"Cinderella attended the royal ball.",
"The prince used the lost glass slipper to search the kingdom.",
"When the slipper fit perfectly, Cinderella was reunited with the prince.",
"Erik Hort's birthplace is Montebello.",
"Marina is bom in Minsk.",
"Montebello is a part of Rockland County."
]
save_dir = 'outputs'# Define save directory for HippoRAG objects (each LLM/Embedding model combination will create a new subdirectory)
llm_model_name = 'gpt-4o-mini' # Any OpenAI model name
embedding_model_name = 'nvidia/NV-Embed-v2'# Embedding model name (NV-Embed, GritLM or Contriever for now)
#Startup a HippoRAG instance
hipporag = HippoRAG(save_dir=save_dir,
llm_model_name=llm_model_name,
embedding_model_name=embedding_model_name)
#Run indexing
hipporag.index(docs=docs)
#Separate Retrieval & QA
queries = [
"What is George Rankin's occupation?",
"How did Cinderella reach her happy ending?",
"What county is Erik Hort's birthplace a part of?"
]
retrieval_results = hipporag.retrieve(queries=queries, num_to_retrieve=2)
qa_results = hipporag.rag_qa(retrieval_results)
#Combined Retrieval & QA
rag_results = hipporag.rag_qa(queries=queries)
#For Evaluation
answers = [
["Politician"],
["By going to the ball."],
["Rockland County"]
]
gold_docs = [
["George Rankin is a politician."],
["Cinderella attended the royal ball.",
"The prince used the lost glass slipper to search the kingdom.",
"When the slipper fit perfectly, Cinderella was reunited with the prince."],
["Erik Hort's birthplace is Montebello.",
"Montebello is a part of Rockland County."]
]
rag_results = hipporag.rag_qa(queries=queries,
gold_docs=gold_docs,
gold_answers=answers)
This simple example will illustrate how to use hipporag
with any vLLM-compatible locally deployed LLM.
- Run a local OpenAI-compatible vLLM server with specified GPUs (make sure you leave enough memory for your embedding model).
export CUDA_VISIBLE_DEVICES=0,1
export VLLM_WORKER_MULTIPROC_METHOD=spawn
export HF_HOME=<path to Huggingface home directory>
conda activate hipporag # vllm should be in this environment
# Tune gpu-memory-utilization or max_model_len to fit your GPU memory, if OOM occurs
vllm serve meta-llama/Llama-3.3-70B-Instruct --tensor-parallel-size 2 --max_model_len 4096 --gpu-memory-utilization 0.95
- Now you can use very similar code to the one above to use
hipporag
:
save_dir = 'outputs'# Define save directory for HippoRAG objects (each LLM/Embedding model combination will create a new subdirectory)
llm_model_name = # Any OpenAI model name
embedding_model_name = # Embedding model name (NV-Embed, GritLM or Contriever for now)
llm_base_url= # Base url for your deployed LLM (i.e. http://localhost:8000/v1)
hipporag = HippoRAG(save_dir=save_dir,
llm_model_name=llm_model,
embedding_model_name=embedding_model_name,
llm_base_url=llm_base_url)
# Same Indexing, Retrieval and QA as running OpenAI models above
To use our code to run experiments we recommend you clone this repository and follow the structure of the main.py
script.
We evaluated several sampled datasets in our paper, some of which are already included in the reproduce/dataset
directory of this repo. For the complete set of datasets, please visit
our HuggingFace dataset and place them under reproduce/dataset
. We also provide the OpenIE results for both gpt-4o-mini
and Llama-3.3-70B-Instruct
for our musique
sample under outputs/musique
.
To test your environment is properly set up, you can use the small dataset reproduce/dataset/sample.json
for debugging as shown below.
Initialize the environmental variables and activate the environment:
export CUDA_VISIBLE_DEVICES=0,1,2,3
export HF_HOME=<path to Huggingface home directory>
export OPENAI_API_KEY=<your openai api key> # if you want to use OpenAI model
conda activate hipporag
dataset=sample # or any other dataset under `reproduce/dataset`
# Run OpenAI model
python main.py --dataset $dataset --llm_base_url https://api.openai.com/v1 --llm_name gpt-4o-mini --embedding_name nvidia/NV-Embed-v2
- As above, run a local OpenAI-compatible vLLM server with specified GPU.
export CUDA_VISIBLE_DEVICES=0,1
export VLLM_WORKER_MULTIPROC_METHOD=spawn
export HF_HOME=<path to Huggingface home directory>
conda activate hipporag # vllm should be in this environment
# Tune gpu-memory-utilization or max_model_len to fit your GPU memory, if OOM occurs
vllm serve meta-llama/Llama-3.3-70B-Instruct --tensor-parallel-size 2 --max_model_len 4096 --gpu-memory-utilization 0.95
- Use another GPUs to run the main program in another terminal.
export CUDA_VISIBLE_DEVICES=2,3 # set another GPUs while vLLM server is running
export HF_HOME=<path to Huggingface home directory>
dataset=sample
python main.py --dataset $dataset --llm_base_url http://localhost:8000/v1 --llm_name meta-llama/Llama-3.3-70B-Instruct --embedding_name nvidia/NV-Embed-v2
vLLM offers an offline batch mode for faster inference, which could bring us more than 3x faster indexing compared to vLLM online server.
- Use the following command to run the main program with vLLM offline batch mode.
export CUDA_VISIBLE_DEVICES=0,1,2,3 # use all GPUs for faster offline indexing
export VLLM_WORKER_MULTIPROC_METHOD=spawn
export HF_HOME=<path to Huggingface home directory>
export OPENAI_API_KEY=''
dataset=sample
python main.py --dataset $dataset --llm_name meta-llama/Llama-3.3-70B-Instruct --openie_mode offline --skip_graph
- After the first step, OpenIE result is saved to file. Go back to run vLLM online server and main program as described in the
Run with vLLM (Llama)
main section.
/reproduce/dataset/sample.json
is a small dataset specifically for debugging.- When debugging vLLM offline mode, set
tensor_parallel_size
as1
inhipporag/llm/vllm_offline.py
. - If you want to rerun a particular experiment, remember to clear the saved files, including OpenIE results and knowledge graph, e.g.,
rm reproduce/dataset/openie_results/openie_sample_results_ner_meta-llama_Llama-3.3-70B-Instruct_3.json
rm -rf outputs/sample/sample_meta-llama_Llama-3.3-70B-Instruct_nvidia_NV-Embed-v2
To setup your own custom dataset for evaluation, follow the format and naming convention shown in reproduce/dataset/sample_corpus.json
(your dataset's name should be followed by _corpus.json
). If running an experiment with pre-defined questions, organize your query corpus according to the query file reproduce/dataset/sample.json
, be sure to also follow our naming convention.
The corpus and optional query JSON files should have the following format:
[
{
"title": "FIRST PASSAGE TITLE",
"text": "FIRST PASSAGE TEXT",
"idx": 0
},
{
"title": "SECOND PASSAGE TITLE",
"text": "SECOND PASSAGE TEXT",
"idx": 1
}
]
[
{
"id": "sample/question_1.json",
"question": "QUESTION",
"answer": [
"ANSWER"
],
"answerable": true,
"paragraphs": [
{
"title": "{FIRST SUPPORTING PASSAGE TITLE}",
"text": "{FIRST SUPPORTING PASSAGE TEXT}",
"is_supporting": true,
"idx": 0
},
{
"title": "{SECOND SUPPORTING PASSAGE TITLE}",
"text": "{SECOND SUPPORTING PASSAGE TEXT}",
"is_supporting": true,
"idx": 1
}
]
}
]
When preparing your data, you may need to chunk each passage, as longer passage may be too complex for the OpenIE process.
π¦ .
β-- π src/hipporag
β βββ π embedding_model # Implementation of all embedding models
β β βββ __init__.py # Getter function for get specific embedding model classes
| | βββ base.py # Base embedding model class `BaseEmbeddingModel` to inherit and `EmbeddingConfig`
| | βββ NVEmbedV2.py # Implementation of NV-Embed-v2 model
| | βββ ...
β βββ π evaluation # Implementation of all evaluation metrics
β β βββ __init__.py
| | βββ base.py # Base evaluation metric class `BaseMetric` to inherit
β β βββ qa_eval.py # Eval metrics for QA
β β βββ retrieval_eval.py # Eval metrics for retrieval
β βββ π information_extraction # Implementation of all information extraction models
β β βββ __init__.py
| | βββ openie_openai_gpt.py # Model for OpenIE with OpenAI GPT
| | βββ openie_vllm_offline.py # Model for OpenIE with LLMs deployed offline with vLLM
β βββ π llm # Classes for inference with large language models
β β βββ __init__.py # Getter function
| | βββ base.py # Config class for LLM inference and base LLM inference class to inherit
| | βββ openai_gpt.py # Class for inference with OpenAI GPT
| | βββ vllm_llama.py # Class for inference using a local vLLM server
| | βββ vllm_offline.py # Class for inference using the vLLM API directly
β βββ π prompts # Prompt templates and prompt template manager class
| β βββ π dspy_prompts # Prompts for filtering
| β β βββ ...
| β βββ π templates # All prompt templates for template manager to load
| β β βββ README.md # Documentations of usage of prompte template manager and prompt template files
| β β βββ __init__.py
| β β βββ triple_extraction.py
| β β βββ ...
β β βββ __init__.py
| | βββ linking.py # Instruction for linking
| | βββ prompt_template_manager.py # Implementation of prompt template manager
β βββ π utils # All utility functions used across this repo (the file name indicates its relevant usage)
β β βββ config_utils.py # We use only one config across all modules and its setup is specified here
| | βββ ...
β βββ __init__.py
β βββ HippoRAG.py # Highest level class for initiating retrieval, question answering, and evaluations
β βββ embedding_store.py # Storage database to load, manage and save embeddings for passages, entities and facts.
β βββ rerank.py # Reranking and filtering methods
β-- π examples
β βββ ...
β βββ ...
β-- π README.md
β-- π requirements.txt # Dependencies list
β-- π .gitignore # Files to exclude from Git
Questions or issues? File an issue or contact Bernal JimΓ©nez GutiΓ©rrez, Yiheng Shu, Yu Su, The Ohio State University
If you find this work useful, please consider citing our papers:
@misc{gutiΓ©rrez2025ragmemorynonparametriccontinual,
title={From RAG to Memory: Non-Parametric Continual Learning for Large Language Models},
author={Bernal JimΓ©nez GutiΓ©rrez and Yiheng Shu and Weijian Qi and Sizhe Zhou and Yu Su},
year={2025},
eprint={2502.14802},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2502.14802},
}
@inproceedings{gutiΓ©rrez2024hipporag,
title={HippoRAG: Neurobiologically Inspired Long-Term Memory for Large Language Models},
author={Bernal JimΓ©nez GutiΓ©rrez and Yiheng Shu and Yu Gu and Michihiro Yasunaga and Yu Su},
booktitle={The Thirty-eighth Annual Conference on Neural Information Processing Systems},
year={2024},
url={https://openreview.net/forum?id=hkujvAPVsg}
- Add support for more embedding models
- Add support for vector database integration
- Add support for embedding endpoints
Please feel free to open an issue or PR if you have any questions or suggestions.