Small yet mighty! 💫

We are releasing SmolVLM: a new 2B small vision language made for on-device use, fine-tunable on consumer GPU, immensely memory efficient 🤠

We release three checkpoints under Apache 2.0: SmolVLM-Instruct, SmolVLM-Synthetic and SmolVLM-Base huggingface.co/collections/...

new loadimg release with more robust url handling
here's an example using fal as an API

Found this little easter egg in one of the blogposts on the HUB.
Apart from efficient upload/download of lfs files this would be my second favorite feature that xet team is working on.
My expectations for the xethub team is off the roof 🔥

Introducing Transformers.js v3: State-of-the-art Machine Learning for the Web. New features include:
⚡ WebGPU support (up to 100x faster than WASM)
🔢 New quantization formats
🏛 121 supported architectures in total
🤖 Over 1200 pre-converted models

Get started with `npm i @huggingface/transformers`

Did you know you can make API requests to VQA models on
huggingface by sending the base64 images in the url field 👀

this allows you to work with local images without the need to upload them online first 🔥

simplified code in the link below
huggingface.co/posts/not-la...

`bitsandbytes` makes it really easy to quantize models

Note: MB should be GB in the diagram.

new SOTA for background removal
beats birefnet 🚀
api ready for personal use 👀
🔗 huggingface.co/spaces/briaa...

got invited to yet another AI event

published a blogpost about integrating custom AI models with huggingface leveraging PyTorchModelHubMixin.
highlights :
* automodel
* custom input types
* trainer api support
* etc , ....
Hope this is informative 🤗
huggingface.co/blog/not-lai...

how come sasha does not have a 🦋 account ?

Another great week in open ML!
Here's a small thread with this week's recap 🧵

💬 Small language models
@huggingface released SmolLM2, a new family of smol language models with Apache 2.0 license with, along with datasets of sizes 135M, 360M and 1.7B 🤗

First post on this app
Happy Halloween (≧∇≦)ﾉ✨