srush / triton-autodiffLinks
Experiment of using Tangent to autodiff triton
☆79Updated last year
Alternatives and similar repositories for triton-autodiff
Users that are interested in triton-autodiff are comparing it to the libraries listed below
Sorting:
- ☆82Updated last year
- A library for unit scaling in PyTorch☆128Updated 3 weeks ago
- Demo of the unit_scaling library, showing how a model can be easily adapted to train in FP8.☆46Updated last year
- ☆113Updated last year
- train with kittens!☆61Updated 9 months ago
- A place to store reusable transformer components of my own creation or found on the interwebs☆59Updated this week
- CUDA and Triton implementations of Flash Attention with SoftmaxN.☆71Updated last year
- An implementation of the Llama architecture, to instruct and delight☆21Updated 2 months ago
- Triton Implementation of HyperAttention Algorithm☆48Updated last year
- Accelerated First Order Parallel Associative Scan☆184Updated 11 months ago
- The simplest implementation of recent Sparse Attention patterns for efficient LLM inference.☆82Updated 2 weeks ago
- Tree Attention: Topology-aware Decoding for Long-Context Attention on GPU clusters☆127Updated 7 months ago
- supporting pytorch FSDP for optimizers☆84Updated 7 months ago
- ☆81Updated last year
- ☆53Updated 9 months ago
- Make triton easier☆47Updated last year
- ☆21Updated 4 months ago
- JAX bindings for Flash Attention v2☆90Updated this week
- A fusion of a linear layer and a cross entropy loss, written for pytorch in triton.☆70Updated last year
- Simple and efficient pytorch-native transformer training and inference (batched)☆78Updated last year
- FlexAttention w/ FlashAttention3 Support☆26Updated 9 months ago
- Automatically take good care of your preemptible TPUs☆36Updated 2 years ago
- A MAD laboratory to improve AI architecture designs 🧪☆123Updated 7 months ago
- A bunch of kernels that might make stuff slower 😉☆56Updated this week
- This repository contains the experimental PyTorch native float8 training UX☆224Updated last year
- ☆53Updated last year
- Load compute kernels from the Hub☆214Updated last week
- The source code of our work "Prepacking: A Simple Method for Fast Prefilling and Increased Throughput in Large Language Models" [AISTATS …☆61Updated 9 months ago
- CUDA implementation of autoregressive linear attention, with all the latest research findings☆44Updated 2 years ago
- Minimal (400 LOC) implementation Maximum (multi-node, FSDP) GPT training☆130Updated last year