deepseek-ai / FlashMLALinks
FlashMLA: Efficient Multi-head Latent Attention Kernels
☆12,438Updated last week
Alternatives and similar repositories for FlashMLA
Users that are interested in FlashMLA are comparing it to the libraries listed below
Sorting:
- DeepEP: an efficient expert-parallel communication library☆8,947Updated this week
- DeepGEMM: clean and efficient FP8 GEMM kernels with fine-grained scaling☆6,138Updated 2 weeks ago
- Production-tested AI infrastructure tools for efficient AGI development and community-driven innovation☆7,961Updated 8 months ago
- A bidirectional pipeline parallelism algorithm for computation-communication overlap in DeepSeek V3/R1 training.☆2,916Updated 2 weeks ago
- Mooncake is the serving platform for Kimi, a leading LLM service provided by Moonshot AI.☆4,632Updated this week
- Expert Parallelism Load Balancer☆1,343Updated 10 months ago
- A Flexible Framework for Experiencing Heterogeneous LLM Inference/Fine-tune Optimizations☆16,406Updated this week
- Analyze computation-communication overlap in V3/R1.☆1,139Updated 10 months ago
- A high-performance distributed file system designed to address the challenges of AI training and inference workloads.☆9,690Updated last week
- DeepSeek-VL2: Mixture-of-Experts Vision-Language Models for Advanced Multimodal Understanding☆5,202Updated 11 months ago
- Domain-specific language designed to streamline the development of high-performance GPU/CPU/Accelerators kernels☆4,863Updated this week
- DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model☆4,990Updated last year
- My learning notes for ML SYS.☆5,157Updated last week
- FlashInfer: Kernel Library for LLM Serving☆4,768Updated this week
- The official repo of MiniMax-Text-01 and MiniMax-VL-01, large-language-model & vision-language-model based on Linear Attention☆3,314Updated 6 months ago
- verl: Volcano Engine Reinforcement Learning for LLMs☆18,756Updated this week
- Lightning-Fast RL for LLM Reasoning and Agents. Made Simple & Flexible.☆3,470Updated this week
- SGLang is a high-performance serving framework for large language models and multimodal models.☆22,800Updated this week
- High-performance inference framework for large language models, focusing on efficiency, flexibility, and availability.☆1,413Updated this week
- Agent framework and applications built upon Qwen>=3.0, featuring Function Calling, MCP, Code Interpreter, RAG, Chrome extension, etc.☆13,053Updated last week
- DeepSeekMoE: Towards Ultimate Expert Specialization in Mixture-of-Experts Language Models☆1,886Updated 2 years ago
- An Easy-to-use, Scalable and High-performance Agentic RL Framework based on Ray (PPO & DAPO & REINFORCE++ & TIS & vLLM & Ray & Async RL)☆8,898Updated last week
- MoBA: Mixture of Block Attention for Long-Context LLMs☆2,038Updated 9 months ago
- Use PEFT or Full-parameter to CPT/SFT/DPO/GRPO 600+ LLMs (Qwen3, Qwen3-MoE, DeepSeek-R1, GLM4.5, InternLM3, Llama4, ...) and 300+ MLLMs (…☆12,337Updated last week
- Fully open reproduction of DeepSeek-R1☆25,848Updated 2 months ago
- A Datacenter Scale Distributed Inference Serving Framework☆6,019Updated this week
- Simple RL training for reasoning☆3,829Updated last month
- Qwen3 is the large language model series developed by Qwen team, Alibaba Cloud.☆26,322Updated 3 weeks ago
- 📚LeetCUDA: Modern CUDA Learn Notes with PyTorch for Beginners🐑, 200+ CUDA Kernels, Tensor Cores, HGEMM, FA-2 MMA.🎉☆9,536Updated 2 weeks ago
- Fast and memory-efficient exact attention☆21,957Updated this week