Hi there 🤗
Long time no see. Summer is now officially over and these last few months have been quite busy at Hugging Face 😎. In this special edition we share some of our most exciting news!
We hope you enjoy it!
Spaces is a simple and free solution to host Machine Learning demo applications using two awesome Python libraries: Gradio and Streamlit. In a matter of minutes you can deploy an app and share it with the community! 🚀
You can sign to the beta today!
New Hub features
In the last few months, the Hub went from 10,000 public model repositories to over 16,000 models! 🚀
Beyond the numbers, we have a ton of other features:
❤️ Like any repo!
🤯 Widgets from object detection to audio classification
↔️ Transferring repos
You can read all about the new features in our special Summer post
Transformers latency down to 1ms? 🤯🤯🤯
We have been working on a really sleek solution to achieve unmatched efficiency for state-of-the-art Transformer models for companies to deploy in their own infrastructure.
⚠️ Join us for a live announcement and demo on Sep 28, where we will be showcasing Infinity for the first time in public!
New in Transformers
Summer has been an exciting time for 🤗 Transformers! The library reached 50,000 stars, 30 million total downloads, and almost 1000 contributors! 🤩
A lot has happened in the last releases:
🖼️🔊New models such as Facebook's DETR for object detection and HuBERT for audio tasks as well as Microsoft's LayoutLMV2 for parsing document images (like PDFs)
💾Integration with the Hub to push your files
🚀and much more!
You can read about all that has happened in our special Summer edition blog post
New in Datasets
You can find 1400 public datasets thanks to the awesome contributions from all our community. 💯
Some of our new features are:
🎧Support in new domains such Automatic Speech Recognition and Image Classification
⬆️Direct hosting and sharing by uploading data files to a repository on the Hub
🔥Dozens of new canonical datasets for everyone to use
AutoNLP In Your Browser
We released a new AutoNLP experience: a web interface to train models straight from your browser!
⚡With just a few clicks:
Now all it takes is a few clicks to train, evaluate and deploy 🤗 Transformers models on your own data.
Try it out - NO CODE needed!
Welcoming new Libraries to the Hub
The Hub is also building great collaborations with Open Source ML libraries to provide free model hosting and versioning!
🔥Sentence Transformers already has over 200 models and you can try them out with new widgets.
✨spaCy shares all its core pipelines in the Hub and now there's a
spacy-huggingface-hub tool to share your pipelines as well!
👥And more integrations with Adapter Transformers, Speechbrain, and other great libraries
The recent TensorBoard integration allows anyone to upload their TensorBoard traces and get an automatic, free instance launched for you. This works with both public and private repositories! We hope this helps during your model development.
A lot more has happened. You can head to our special summer edition blog post to read all about it.
👋 Until next time!