Lists (1)
Stars
Code for the preprint "Cache Me If You Can: How Many KVs Do You Need for Effective Long-Context LMs?"
[ISCA'25] LIA: A Single-GPU LLM Inference Acceleration with Cooperative AMX-Enabled CPU-GPU Computation and CXL Offloading
Efficient Compute-Communication Overlap for Distributed LLM Inference
slime is a LLM post-training framework aiming at scaling RL.
Repo for SeedVR2 & SeedVR (CVPR2025 Highlight)
TreeRL: LLM Reinforcement Learning with On-Policy Tree Search in ACL'25
MiniMax-M1, the world's first open-weight, large-scale hybrid-attention reasoning model.
各地房价热力图，杭州、北京、上海、苏州、天津、成都、南京、长沙、无锡、南宁、太原、青岛、南昌、郑州
A list of works on video generation towards world model
Query-agnostic KV cache eviction: 3–4× reduction in memory and 2× decrease in latency (Qwen3/2.5, Gemma3, LLaMA3)
The repository for ATC'25 paper "Greyhound: Hunting Fail-Slows in Hybrid-Parallel Training at Scale"
[NeurIPS 2024] Code for the paper "Diffusion of Thoughts: Chain-of-Thought Reasoning in Diffusion Language Models"