haoliuhl / ringattentionLinks
Large Context Attention
β714Updated 4 months ago
Alternatives and similar repositories for ringattention
Users that are interested in ringattention are comparing it to the libraries listed below
Sorting:
- Implementation of π Ring Attention, from Liu et al. at Berkeley AI, in Pytorchβ514Updated 3 weeks ago
- Ring attention implementation with flash attentionβ774Updated 2 weeks ago
- [ICML 2024] Break the Sequential Dependency of LLM Inference Using Lookahead Decodingβ1,249Updated 2 months ago
- [ICLR 2024] Sheared LLaMA: Accelerating Language Model Pre-training via Structured Pruningβ612Updated last year
- Microsoft Automatic Mixed Precision Libraryβ602Updated 8 months ago
- [ICML 2024] CLLMs: Consistency Large Language Modelsβ391Updated 6 months ago
- Memory optimization and training recipes to extrapolate language models' context length to 1 million tokens, with minimal hardware.β727Updated 8 months ago
- π Efficiently (pre)training foundation models with native PyTorch features, including FSDP for training and SDPA implementation of Flashβ¦β249Updated this week
- Official repository for DistFlashAttn: Distributed Memory-efficient Attention for Long-context LLMs Trainingβ210Updated 9 months ago
- Triton-based implementation of Sparse Mixture of Experts.β217Updated 6 months ago
- Explorations into some recent techniques surrounding speculative decodingβ268Updated 5 months ago
- USP: Unified (a.k.a. Hybrid, 2D) Sequence Parallel Attention for Long Context Transformers Model Training and Inferenceβ510Updated last week
- [NeurIPS 2024] KVQuant: Towards 10 Million Context Length LLM Inference with KV Cache Quantizationβ357Updated 9 months ago
- Helpful tools and examples for working with flex-attentionβ811Updated this week
- β222Updated 11 months ago
- This repository contains the experimental PyTorch native float8 training UXβ223Updated 10 months ago
- Scalable toolkit for efficient model alignmentβ807Updated this week
- distributed trainer for LLMsβ575Updated last year
- β144Updated 2 years ago
- Pipeline Parallelism for PyTorchβ767Updated 9 months ago
- β322Updated last year
- A subset of PyTorch's neural network modules, written in Python using OpenAI's Triton.β546Updated this week
- Zero Bubble Pipeline Parallelismβ396Updated 3 weeks ago
- LLM KV cache compression made easyβ493Updated 3 weeks ago
- β456Updated last week
- Tutel MoE: Optimized Mixture-of-Experts Library, Support DeepSeek FP8/FP4β829Updated this week
- FP16xINT4 LLM inference kernel that can achieve near-ideal ~4x speedups up to medium batchsizes of 16-32 tokens.β831Updated 9 months ago
- [ICLR 2025] DuoAttention: Efficient Long-Context LLM Inference with Retrieval and Streaming Headsβ463Updated 3 months ago
- For releasing code related to compression methods for transformers, accompanying our publicationsβ429Updated 4 months ago
- Code for the paper "QMoE: Practical Sub-1-Bit Compression of Trillion-Parameter Models".β275Updated last year