graphcore-research / out-of-the-box-fp8-trainingLinks
Demo of the unit_scaling library, showing how a model can be easily adapted to train in FP8.
☆45Updated last year
Alternatives and similar repositories for out-of-the-box-fp8-training
Users that are interested in out-of-the-box-fp8-training are comparing it to the libraries listed below
Sorting:
- Experiment of using Tangent to autodiff triton☆80Updated last year
- ☆91Updated last year
- A place to store reusable transformer components of my own creation or found on the interwebs☆60Updated last week
- CUDA implementation of autoregressive linear attention, with all the latest research findings☆45Updated 2 years ago
- A library for unit scaling in PyTorch☆130Updated 3 months ago
- Personal solutions to the Triton Puzzles☆20Updated last year
- ☆121Updated last year
- This repository contains the experimental PyTorch native float8 training UX☆223Updated last year
- PyTorch centric eager mode debugger☆48Updated 10 months ago
- FlexAttention w/ FlashAttention3 Support☆27Updated last year
- Make triton easier☆48Updated last year
- A bunch of kernels that might make stuff slower 😉☆62Updated this week
- CUDA and Triton implementations of Flash Attention with SoftmaxN.☆73Updated last year
- Hacks for PyTorch☆19Updated 2 years ago
- train with kittens!☆63Updated 11 months ago
- ☆21Updated 7 months ago
- Triton Implementation of HyperAttention Algorithm☆48Updated last year
- ☆58Updated last year
- Tree Attention: Topology-aware Decoding for Long-Context Attention on GPU clusters☆130Updated 10 months ago
- Minimal (400 LOC) implementation Maximum (multi-node, FSDP) GPT training☆132Updated last year
- ☆28Updated 9 months ago
- JAX bindings for Flash Attention v2☆97Updated this week
- Transformer with Mu-Parameterization, implemented in Jax/Flax. Supports FSDP on TPU pods.☆32Updated 4 months ago
- ☆112Updated last year
- ☆53Updated last year
- Utilities for Training Very Large Models☆58Updated last year
- ☆149Updated 2 years ago
- ☆83Updated last year
- Automatically take good care of your preemptible TPUs☆37Updated 2 years ago
- ☆46Updated last year