lucidrains / flash-attention-jaxLinks
Implementation of Flash Attention in Jax
☆213Updated last year
Alternatives and similar repositories for flash-attention-jax
Users that are interested in flash-attention-jax are comparing it to the libraries listed below
Sorting:
- jax-triton contains integrations between JAX and OpenAI Triton☆400Updated 2 weeks ago
- Implementation of a Transformer, but completely in Triton☆268Updated 3 years ago
- ☆186Updated 2 weeks ago
- JAX Synergistic Memory Inspector☆174Updated 11 months ago
- JMP is a Mixed Precision library for JAX.☆202Updated 4 months ago
- JAX implementation of the Llama 2 model☆218Updated last year
- JAX bindings for Flash Attention v2☆89Updated 11 months ago
- A library for unit scaling in PyTorch☆125Updated 6 months ago
- ☆355Updated last year
- A simple library for scaling up JAX programs☆139Updated 7 months ago
- LoRA for arbitrary JAX models and functions☆138Updated last year
- Implementation of the specific Transformer architecture from PaLM - Scaling Language Modeling with Pathways - in Jax (Equinox framework)☆187Updated 2 years ago
- Accelerated First Order Parallel Associative Scan☆182Updated 10 months ago
- ☆317Updated this week
- Implementation of fused cosine similarity attention in the same style as Flash Attention☆214Updated 2 years ago
- This repository contains the experimental PyTorch native float8 training UX☆224Updated 10 months ago
- ☆60Updated 3 years ago
- Train very large language models in Jax.☆204Updated last year
- Memory Efficient Attention (O(sqrt(n)) for Jax and PyTorch☆184Updated 2 years ago
- Pax is a Jax-based machine learning framework for training large scale models. Pax allows for advanced and fully configurable experimenta…☆508Updated last week
- ☆108Updated last year
- Inference code for LLaMA models in JAX☆118Updated last year
- ☆67Updated 2 years ago
- Named tensors with first-class dimensions for PyTorch☆331Updated 2 years ago
- Experiment of using Tangent to autodiff triton☆79Updated last year
- Implementation of a memory efficient multi-head attention as proposed in the paper, "Self-attention Does Not Need O(n²) Memory"☆379Updated last year
- CLU lets you write beautiful training loops in JAX.☆345Updated this week
- Orbax provides common checkpointing and persistence utilities for JAX users☆390Updated this week
- Implementation of Mega, the Single-head Attention with Multi-headed EMA architecture that currently holds SOTA on Long Range Arena☆204Updated last year
- seqax = sequence modeling + JAX☆159Updated last week