subaru kimura

joddy

AI & ML interests

None yet

Recent Activity

reacted to m-ric's post with 👍 15 days ago

𝐇𝐮𝐠𝐠𝐢𝐧𝐠 𝐅𝐚𝐜𝐞 𝐫𝐞𝐥𝐞𝐚𝐬𝐞𝐬 𝐏𝐢𝐜𝐨𝐭𝐫𝐨𝐧, 𝐚 𝐦𝐢𝐜𝐫𝐨𝐬𝐜𝐨𝐩𝐢𝐜 𝐥𝐢𝐛 𝐭𝐡𝐚𝐭 𝐬𝐨𝐥𝐯𝐞𝐬 𝐋𝐋𝐌 𝐭𝐫𝐚𝐢𝐧𝐢𝐧𝐠 𝟒𝐃 𝐩𝐚𝐫𝐚𝐥𝐥𝐞𝐥𝐢𝐳𝐚𝐭𝐢𝐨𝐧 🥳 🕰️ Llama-3.1-405B took 39 million GPU-hours to train, i.e. about 4.5 thousand years. 👴🏻 If they had needed all this time, we would have GPU stories from the time of Pharaoh 𓂀: "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 " 🛠️ 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. 🤏 𝗕𝘂𝘁 𝗻𝗼𝘄 𝘄𝗲 𝗱𝗼𝗻'𝘁 𝗻𝗲𝗲𝗱 𝗵𝘂𝗴𝗲 𝗿𝗲𝗽𝗼𝘀 𝗮𝗻𝘆𝗺𝗼𝗿𝗲! 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! ⚡ 𝗜𝘁'𝘀 𝘁𝗶𝗻𝘆, 𝘆𝗲𝘁 𝗽𝗼𝘄𝗲𝗿𝗳𝘂𝗹: 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 👉 https://github.com/huggingface/picotron/tree/main/picotron

published a model 2 months ago

joddy/sft_output

liked a model 2 months ago

microsoft/phi-4

View all activity

Organizations

joddy's activity

upvoted 4 collections 6 months ago

upvoted a paper 7 months ago

Windows Agent Arena: Evaluating Multi-Modal OS Agents at Scale

Paper • 2409.08264 • Published Sep 12, 2024 • 46

upvoted 3 collections 8 months ago

LLaVA-Next-Interleave

Collection

7 items • Updated Oct 4, 2024 • 16

LLaVA-Video

Collection

Models focus on video understanding (previously known as LLaVA-NeXT-Video). • 8 items • Updated Feb 21 • 61

VILA: On Pre-training for Visual Language Models

Collection

10 items • Updated Oct 31, 2024 • 53

upvoted a collection 10 months ago

PaliGemma Release

Collection

Pretrained and mix checkpoints for PaliGemma • 16 items • Updated about 5 hours ago • 145

upvoted a collection 11 months ago

Vision Language Models Papers 🖼️💬📝

Collection

Papers about vision-language models, most important ones are on top of the list. • 27 items • Updated Apr 30, 2024 • 36