Aurรฉlien-Morgan CLAUDON

Aurelien-Morgan

AI & ML interests

None yet

Recent Activity

liked a Space about 12 hours ago
akhaliq/anychat
reacted to m-ric's post with ๐Ÿ‘ 1 day ago
๐‡๐ฎ๐ ๐ ๐ข๐ง๐  ๐…๐š๐œ๐ž ๐ซ๐ž๐ฅ๐ž๐š๐ฌ๐ž๐ฌ ๐๐ข๐œ๐จ๐ญ๐ซ๐จ๐ง, ๐š ๐ฆ๐ข๐œ๐ซ๐จ๐ฌ๐œ๐จ๐ฉ๐ข๐œ ๐ฅ๐ข๐› ๐ญ๐ก๐š๐ญ ๐ฌ๐จ๐ฅ๐ฏ๐ž๐ฌ ๐‹๐‹๐Œ ๐ญ๐ซ๐š๐ข๐ง๐ข๐ง๐  ๐Ÿ’๐ƒ ๐ฉ๐š๐ซ๐š๐ฅ๐ฅ๐ž๐ฅ๐ข๐ณ๐š๐ญ๐ข๐จ๐ง ๐Ÿฅณ ๐Ÿ•ฐ๏ธ Llama-3.1-405B took 39 million GPU-hours to train, i.e. about 4.5 thousand years. ๐Ÿ‘ด๐Ÿป If they had needed all this time, we would have GPU stories from the time of Pharaoh ๐“‚€: "Alas, Lord of Two Lands, the shipment of counting-stones arriving from Cathay was lost to pirates, this shall delay the building of your computing temple by many moons " ๐Ÿ› ๏ธ But instead, they just parallelized the training on 24k H100s, which made it take just a few months. This required parallelizing across 4 dimensions: data, tensor, context, pipeline. And it is infamously hard to do, making for bloated code repos that hold together only by magic. ๐Ÿค ๐—•๐˜‚๐˜ ๐—ป๐—ผ๐˜„ ๐˜„๐—ฒ ๐—ฑ๐—ผ๐—ป'๐˜ ๐—ป๐—ฒ๐—ฒ๐—ฑ ๐—ต๐˜‚๐—ด๐—ฒ ๐—ฟ๐—ฒ๐—ฝ๐—ผ๐˜€ ๐—ฎ๐—ป๐˜†๐—บ๐—ผ๐—ฟ๐—ฒ! Instead of building mega-training codes, Hugging Face colleagues cooked in the other direction, towards tiny 4D parallelism libs. A team has built Nanotron, already widely used in industry. And now a team releases Picotron, a radical approach to code 4D Parallelism in just a few hundred lines of code, a real engineering prowess, making it much easier to understand what's actually happening! โšก ๐—œ๐˜'๐˜€ ๐˜๐—ถ๐—ป๐˜†, ๐˜†๐—ฒ๐˜ ๐—ฝ๐—ผ๐˜„๐—ฒ๐—ฟ๐—ณ๐˜‚๐—น: Counting in MFU (Model FLOPs Utilization, how much the model actually uses all the compute potential), this lib reaches ~50% on SmolLM-1.7B model with 8 H100 GPUs, which is really close to what huge libs would reach. (Caution: the team is leading further benchmarks to verify this) Go take a look ๐Ÿ‘‰ https://github.com/huggingface/picotron/tree/main/picotron
View all activity

Articles

Organizations

Giskard's profile picture Gradio-Blocks-Party's profile picture Keras Dreambooth Event's profile picture Blog-explorers's profile picture huggingPartyParis's profile picture ZeroGPU Explorers's profile picture C4AI Community's profile picture Chinese LLMs on Hugging Face's profile picture Paris AI Running Club's profile picture cvmistralparis's profile picture Hugging Face Discord Community's profile picture Hugging Face Party @ PyTorch Conference's profile picture Nerdy Face's profile picture retrain-pipelines's profile picture

Aurelien-Morgan's activity

upvoted an article 24 days ago
view article
Article

Letโ€™s make a generation of amazing image generation models

By burtenshaw โ€ข
โ€ข 33
upvoted 3 articles 2 months ago
view article
Article

Fixing Gradient Accumulation

โ€ข 43
view article
Article

Model Card Generator Interface: Crafting Clear Insights into AI Models

By mitalipo โ€ข
โ€ข 4
upvoted an article 2 months ago
view article
Article

A Short Summary of Chinese AI Global Expansion

โ€ข 19
upvoted an article 3 months ago
view article
Article

Llama can now see and run on your device - welcome Llama 3.2

โ€ข 180
upvoted 3 articles 3 months ago
view article
Article

Introducing Community Tools on HuggingChat

โ€ข 34
view article
Article

The Environmental Impacts of AI -- Primer

By sasha โ€ข
โ€ข 33