Infiniband Verbs Performance Tests

Virtualized Elastic KV Cache for Dynamic GPU Sharing and Beyond

My learning notes/codes for ML SYS.

Checkpoint-engine is a simple middleware to update model weights in LLM inference engines

ScreenCoder — Turn any UI screenshot into clean, editable HTML/CSS with full control. Fast, accurate, and easy to customize.

An Efficient and User-Friendly Scaling Library for Reinforcement Learning with Large Language Models

[arXiv 2025] MMSI-Bench: A Benchmark for Multi-Image Spatial Intelligence

Multi-SpatialMLLM Multi-Frame Spatial Understanding with Multi-Modal Large Language Models

DeepSeek-V3/R1 inference performance simulator

A high-performance distributed file system designed to address the challenges of AI training and inference workloads.

Domain-specific language designed to streamline the development of high-performance GPU/CPU/Accelerators kernels

DeepGEMM: clean and efficient FP8 GEMM kernels with fine-grained scaling

DeepEP: an efficient expert-parallel communication library

FlashMLA: Efficient Multi-head Latent Attention Kernels

MLGym A New Framework and Benchmark for Advancing AI Research Agents

A scalable generative AI framework built for researchers and developers working on Large Language Models, Multimodal, and Speech AI (Automatic Speech Recognition and Text-to-Speech)

A bibliography and survey of the papers surrounding o1

[NeurIPS 2024] Efficient LLM Scheduling by Learning to Rank

FlexFlow Serve: Low-Latency, High-Performance LLM Serving

A unified inference and post-training framework for accelerated video generation.

[ICLR 2025] COAT: Compressing Optimizer States and Activation for Memory-Efficient FP8 Training

https://wavespeed.ai/ Context parallel attention that accelerates DiT model inference with dynamic caching

verl: Volcano Engine Reinforcement Learning for LLMs

[COLM 2024] SKVQ: Sliding-window Key and Value Cache Quantization for Large Language Models

[CoRL 2024] VLM-Grounder: A VLM Agent for Zero-Shot 3D Visual Grounding

制作懂人情世故的大语言模型 | 涵盖提示词工程、RAG、Agent、LLM微调教程

A throughput-oriented high-performance serving framework for LLMs