verl: Volcano Engine Reinforcement Learning for LLMs

[NeurIPS'24 Spotlight, ICLR'25] To speed up Long-context LLMs' inference, approximate and dynamic sparse calculate the attention, which reduces inference latency by up to 10x for pre-filling on an …

Skyrise is a research project exploring data processing on elastic cloud resources.

EuroSys '24: "Trinity: A Fast Compressed Multi-attribute Data Store"

[ICLR2025 Spotlight] MagicPIG: LSH Sampling for Efficient LLM Generation

10x Faster Long-Context LLM By Smart KV Cache Optimizations

OpenR: An Open Source Framework for Advanced Reasoning with Large Language Models

TAG-Bench: A benchmark for table-augmented generation (TAG)

A throughput-oriented high-performance serving framework for LLMs

Python-based research interface for blackbox and hyperparameter optimization, based on the internal Google Vizier Service.

A Flexible Framework for Experiencing Cutting-edge LLM Inference Optimizations

Generative AI extensions for onnxruntime

[EMNLP 2022] An Open Toolkit for Knowledge Graph Extraction and Construction

The code used to train and run inference with the ColVision models, e.g. ColPali, ColQwen2, and ColSmol.

GNN-RAG: Graph Neural Retrieval for Large Language Modeling Reasoning

OpenCompass is an LLM evaluation platform, supporting a wide range of models (Llama3, Mistral, InternLM2,GPT-4,LLaMa2, Qwen,GLM, Claude, etc) over 100+ datasets.

StaRD: Statute Retrieval Dataset based on Real-World Legal Consultation

Minimal keyword extraction with BERT

Faster Learned Sparse Retrieval with Block-Max Pruning. ACM SIGIR 2024.

Official Implementation of "Multi-Head RAG: Solving Multi-Aspect Problems with LLMs"

Using GPT to organize and access information, and generate questions. Long term goal is to make an agent-like research assistant.

[Paper List] Papers integrating knowledge graphs (KGs) and large language models (LLMs)

Official Implementation of NeurIPS 2024 paper "G-Retriever: Retrieval-Augmented Generation for Textual Graph Understanding and Question Answering""

State-of-the-Art Text Embeddings

Daichi Amagata, ICDE2024