li-plus / flash-preferenceLinks
Accelerate LLM preference tuning via prefix sharing with a single line of code
☆51Updated 4 months ago
Alternatives and similar repositories for flash-preference
Users that are interested in flash-preference are comparing it to the libraries listed below
Sorting:
- Quantized Attention on GPU☆44Updated 11 months ago
- Odysseus: Playground of LLM Sequence Parallelism☆78Updated last year
- ☆50Updated 5 months ago
- 🔥 LLM-powered GPU kernel synthesis: Train models to convert PyTorch ops into optimized Triton kernels via SFT+RL. Multi-turn compilation…☆98Updated this week
- ☆130Updated 5 months ago
- ☆106Updated 5 months ago
- Estimate MFU for DeepSeekV3☆26Updated 10 months ago
- An efficient implementation of the NSA (Native Sparse Attention) kernel☆124Updated 4 months ago
- A Suite for Parallel Inference of Diffusion Transformers (DiTs) on multi-GPU Clusters☆51Updated last year
- ☆39Updated 3 months ago
- A lightweight reinforcement learning framework that integrates seamlessly into your codebase, empowering developers to focus on algorithm…☆75Updated 2 months ago
- 16-fold memory access reduction with nearly no loss☆106Updated 7 months ago
- [ACL 2024] RelayAttention for Efficient Large Language Model Serving with Long System Prompts☆40Updated last year
- Flash-Muon: An Efficient Implementation of Muon Optimizer☆206Updated 5 months ago
- ☆147Updated 9 months ago
- The evaluation framework for training-free sparse attention in LLMs☆102Updated last month
- [ICLR2025] Breaking Throughput-Latency Trade-off for Long Sequences with Speculative Decoding☆131Updated 11 months ago
- Transformers components but in Triton☆34Updated 6 months ago
- Code for paper: [ICLR2025 Oral] FlexPrefill: A Context-Aware Sparse Attention Mechanism for Efficient Long-Sequence Inference☆152Updated last month
- ☆65Updated 6 months ago
- ☆120Updated 2 months ago
- Fast and memory-efficient exact attention☆14Updated this week
- DeeperGEMM: crazy optimized version☆73Updated 6 months ago
- ☆72Updated 3 weeks ago
- The official implementation of paper: SimLayerKV: A Simple Framework for Layer-Level KV Cache Reduction.☆50Updated last year
- Beyond KV Caching: Shared Attention for Efficient LLMs☆20Updated last year
- Framework to reduce autotune overhead to zero for well known deployments.☆85Updated last month
- ☆96Updated 8 months ago
- Implementation for FP8/INT8 Rollout for RL training without performence drop.☆269Updated last week
- Xmixers: A collection of SOTA efficient token/channel mixers☆29Updated 2 months ago