RulinShao / LightSeqLinks
Official repository for DistFlashAttn: Distributed Memory-efficient Attention for Long-context LLMs Training
☆214Updated last year
Alternatives and similar repositories for LightSeq
Users that are interested in LightSeq are comparing it to the libraries listed below
Sorting:
- Triton-based implementation of Sparse Mixture of Experts.☆233Updated 8 months ago
- Best practices for training DeepSeek, Mixtral, Qwen and other MoE models using Megatron Core.☆56Updated this week
- PyTorch bindings for CUTLASS grouped GEMM.☆109Updated 2 months ago
- ☆110Updated 11 months ago
- PyTorch bindings for CUTLASS grouped GEMM.☆135Updated last month
- ☆273Updated last month
- GEAR: An Efficient KV Cache Compression Recipefor Near-Lossless Generative Inference of LLM☆165Updated last year
- ☆332Updated last year
- ☆85Updated 3 years ago
- [NeurIPS 2024] KVQuant: Towards 10 Million Context Length LLM Inference with KV Cache Quantization☆372Updated last year
- [ICML 2024] Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference☆321Updated last month
- [ICLR2025] Breaking Throughput-Latency Trade-off for Long Sequences with Speculative Decoding☆124Updated 8 months ago
- [ICLR 2025] COAT: Compressing Optimizer States and Activation for Memory-Efficient FP8 Training☆227Updated 2 weeks ago
- REST: Retrieval-Based Speculative Decoding, NAACL 2024☆207Updated 8 months ago
- ☆139Updated 6 months ago
- [MLSys'24] Atom: Low-bit Quantization for Efficient and Accurate LLM Serving☆319Updated last year
- Ouroboros: Speculative Decoding with Large Model Enhanced Drafting (EMNLP 2024 main)☆109Updated 5 months ago
- [NeurIPS 2024] The official implementation of "Kangaroo: Lossless Self-Speculative Decoding for Accelerating LLMs via Double Early Exitin…☆59Updated last year
- 16-fold memory access reduction with nearly no loss☆104Updated 4 months ago
- Bridge Megatron-Core to Hugging Face/Reinforcement Learning☆93Updated last week
- Odysseus: Playground of LLM Sequence Parallelism☆76Updated last year
- 🚀 Collection of components for development, training, tuning, and inference of foundation models leveraging PyTorch native components.☆208Updated last week
- Boosting 4-bit inference kernels with 2:4 Sparsity☆80Updated 11 months ago
- Zero Bubble Pipeline Parallelism☆418Updated 3 months ago
- ☆123Updated 2 months ago
- Code associated with the paper **Draft & Verify: Lossless Large Language Model Acceleration via Self-Speculative Decoding**☆201Updated 6 months ago
- Explorations into some recent techniques surrounding speculative decoding☆282Updated 8 months ago
- PyTorch implementation of paper "Response Length Perception and Sequence Scheduling: An LLM-Empowered LLM Inference Pipeline".☆90Updated 2 years ago
- [NeurIPS'23] H2O: Heavy-Hitter Oracle for Efficient Generative Inference of Large Language Models.☆466Updated last year
- Flash-LLM: Enabling Cost-Effective and Highly-Efficient Large Generative Model Inference with Unstructured Sparsity☆217Updated last year