Loading video...

Video Failed to Load

Go Home

Introducing SmolVLM 256M (& 500M): The world’s smallest multimodal model. Designed for efficiency and perfect for on-device applications! 🔥 It’s so small, it can even run 100% locally in your browser on WebGPU! 🤏 Powered by Transformers.js! ⚡️ Try it out yourself! 👇

25,302 views • 1 year ago •via X (Twitter)

11 Comments

Xenova's profile picture
Xenova1 year ago

Probably the most impressive thing is that both models (256M & 500M) surpass the performance of Idefics 80B (released less than 1.5 years ago). 🤯 Open source & open science FTW! 🔥 Read our blog post to learn more:

Xenova's profile picture
Xenova1 year ago

Important links: - Collection (models & demos): - WebGPU demo:

WillScuderi's profile picture
WillScuderi1 year ago

🚀 We've just launched the ultimate travel companion for all business travellers. The espresso Display 15 with Stand+ features: ✅ Aluminium Build ✅ Works with one cable for Mac, PC, iphone (15 and later) ✅ Display above laptop ✅ 1080p/16m colours/300nits ✅ 0.2"/5mm thin

tkmadeit's profile picture
tkmadeit1 year ago

Was really hoping it would ace this

Moes's profile picture
Moes1 year ago

Wait. Waittttt. Could you make it use computer?

James's profile picture
James1 year ago

is there a notebook somewhere to finetune this?

Calvin Muchemwa's profile picture
Calvin Muchemwa1 year ago

W

ZOHEB's profile picture
ZOHEB1 year ago

Small models for the win!

MacroHard's profile picture
MacroHard1 year ago

Could this model perform a type of analysis on a given picture? I'm thinking like a picture of a half eaten sheet cake (or w/e lol) the model can return an eyeball percentage like a human could?

Harry 🐺's profile picture
Harry 🐺1 year ago

Keep up the good work mate! These are fantastic!

tz 🏴‍☠️'s profile picture
tz 🏴‍☠️1 year ago

fantastic

Related Videos