haoliuhl / ringattention
Transformers with Arbitrarily Large Context
β641Updated 3 months ago
Related projects β
Alternatives and complementary repositories for ringattention
- Implementation of π Ring Attention, from Liu et al. at Berkeley AI, in Pytorchβ476Updated 3 weeks ago
- Ring attention implementation with flash attentionβ585Updated last week
- Memory optimization and training recipes to extrapolate language models' context length to 1 million tokens, with minimal hardware.β647Updated last month
- Scalable toolkit for efficient model alignmentβ620Updated this week
- π Efficiently (pre)training foundation models with native PyTorch features, including FSDP for training and SDPA implementation of Flashβ¦β193Updated this week
- Official repository for LightSeq: Sequence Level Parallelism for Distributed Training of Long Context Transformersβ195Updated 3 months ago
- Microsoft Automatic Mixed Precision Libraryβ525Updated last month
- [ICLR 2024] Sheared LLaMA: Accelerating Language Model Pre-training via Structured Pruningβ558Updated 8 months ago
- Helpful tools and examples for working with flex-attentionβ469Updated 3 weeks ago
- Explorations into some recent techniques surrounding speculative decodingβ211Updated last year
- USP: Unified (a.k.a. Hybrid, 2D) Sequence Parallel Attention for Long Context Transformers Model Training and Inferenceβ357Updated this week
- [ICML 2024] Break the Sequential Dependency of LLM Inference Using Lookahead Decodingβ1,149Updated last month
- [NeurIPS 2024] KVQuant: Towards 10 Million Context Length LLM Inference with KV Cache Quantizationβ305Updated 3 months ago
- For releasing code related to compression methods for transformers, accompanying our publicationsβ372Updated last month
- Official PyTorch implementation of QA-LoRAβ117Updated 8 months ago
- Zero Bubble Pipeline Parallelismβ281Updated last week
- Pipeline Parallelism for PyTorchβ726Updated 2 months ago
- Triton-based implementation of Sparse Mixture of Experts.β185Updated last month
- Implementation of ST-Moe, the latest incarnation of MoE after years of research at Brain, in Pytorchβ293Updated 5 months ago
- [ICML 2024] CLLMs: Consistency Large Language Modelsβ353Updated this week
- This repository contains the experimental PyTorch native float8 training UXβ211Updated 3 months ago
- Code for paper: "QuIP: 2-Bit Quantization of Large Language Models With Guarantees"β350Updated 8 months ago
- Scaling Data-Constrained Language Modelsβ321Updated last month
- batched lorasβ336Updated last year
- [ICML 2024] SqueezeLLM: Dense-and-Sparse Quantizationβ649Updated 3 months ago
- A simple and effective LLM pruning approach.β669Updated 3 months ago
- β289Updated 7 months ago
- GPTQ inference Triton kernelβ284Updated last year
- Code for the paper "QMoE: Practical Sub-1-Bit Compression of Trillion-Parameter Models".β262Updated last year