Share this page!

Enter URL or ID to Unroll

You can paste full URL like: https://x.com/threadreaderapp/status/1644127596119195649
or just the ID like: 1644127596119195649

How to get URL link on X (Twitter) App

  1. On the Twitter thread, click on or icon on the bottom
  2. Click again on or Share Via icon
  3. Click on Copy Link to Tweet
  4. Paste it above and click "Unroll Thread"!
  5. More info at Twitter Help
Tri Dao Profile picture
Tri Dao
@tri_dao
Incoming Asst. Prof @PrincetonCS, Chief Scientist @togethercompute. Machine learning & systems.
Jerome Ku Profile picture 1 subscribed
Jul 17, 2023 • 13 tweets • 4 min read
Announcing FlashAttention-2! We released FlashAttention a year ago, making attn 2-4 faster and is now widely used in most LLM libraries. Recently I’ve been working on the next version: 2x faster than v1, 5-9x vs standard attn, reaching 225 TFLOPs/s training speed on A100. 1/
Image
Image The tech report has all the info:

More details in blogposts:
https://t.co/hh2yGicgOe
https://t.co/ANwdH0fgMs
https://t.co/EjeYlGmBuL

FlashAttention-2 is available in the open source: https://t.co/b3RaWgoFbE
2/tridao.me/publications/f…
crfm.stanford.edu/2023/07/17/fla…
together.ai/blog/tri-dao-f…
princeton-nlp.github.io/flash-atttenti…
github.com/Dao-AILab/flas…
Jan 17, 2023 • 8 tweets • 5 min read
I’ve been working with @AdeptAILabs and we’ve made FlashAttention even faster for long sequences! For seqlen 8K, FlashAttention is now up to 2.7x faster than a standard PyTorch implementation even at small batch, making it easier to train better LMs with longer context 1/7 Image We described this improvement in more details in this blogpost:
crfm.stanford.edu/2023/01/13/fla…
adept.ai/flashierattent…
2/7
Nov 29, 2022 • 6 tweets • 2 min read
We're releasing an optimized implementation of GPT2/GPT3 with FlashAttention🚀!
This trains 3-5x faster than the Huggingface version, reaching up to 189 TFLOPs/sec per A100, 60.6% (model) FLOPs util of the theoretical maximum. 1/6
github.com/HazyResearch/f… The main ingredient is FlashAttention, which computes attention fast (2-4x) and with less memory (10x), without any approximation. This means that we don't need to do any activation checkpointing
2/6
May 31, 2022 • 11 tweets • 5 min read
Announcing FlashAttention, a fast and memory-efficient attention algorithm with no approximation! 📣 w/ @realDanFu

By reducing GPU memory reads/writes, FlashAttention runs 2-4x faster & requires 5-20x less memory than PyTorch standard attention, & scales to seq. length 64K. 1/ Image Transformers have grown larger and deeper, but longer context remains difficult, since self-attention has time and memory quadratic in seq. length. Approx attn attempts to address this by trading off quality for compute complexity, but often doesn’t achieve wall-clock speedup. 2/