thu-nics / MoALinks
The official implementation of the paper <MoA: Mixture of Sparse Attention for Automatic Large Language Model Compression>
☆134Updated 3 weeks ago
Alternatives and similar repositories for MoA
Users that are interested in MoA are comparing it to the libraries listed below
Sorting:
- Code for paper: [ICLR2025 Oral] FlexPrefill: A Context-Aware Sparse Attention Mechanism for Efficient Long-Sequence Inference☆113Updated last month
- SeerAttention: Learning Intrinsic Sparse Attention in Your LLMs☆128Updated this week
- 16-fold memory access reduction with nearly no loss☆99Updated 2 months ago
- XAttention: Block Sparse Attention with Antidiagonal Scoring☆166Updated this week
- [ICML24] Pruner-Zero: Evolving Symbolic Pruning Metric from scratch for LLMs☆83Updated 7 months ago
- ☆130Updated 4 months ago
- GEAR: An Efficient KV Cache Compression Recipefor Near-Lossless Generative Inference of LLM☆163Updated 11 months ago
- [ACL 2024] Not All Experts are Equal: Efficient Expert Pruning and Skipping for Mixture-of-Experts Large Language Models☆92Updated last year
- ☆104Updated 2 weeks ago
- ☆114Updated 3 weeks ago
- Efficient triton implementation of Native Sparse Attention.☆168Updated last month
- Official implementation of "Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decoding"☆233Updated 2 weeks ago
- ☆256Updated last year
- [ICML 2024] Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference☆297Updated 7 months ago
- The official implementation of the paper "Towards Efficient Mixture of Experts: A Holistic Study of Compression Techniques (TMLR)".☆71Updated 3 months ago
- A sparse attention kernel supporting mix sparse patterns☆238Updated 4 months ago
- An algorithm for weight-activation quantization (W4A4, W4A8) of LLMs, supporting both static and dynamic quantization