OPEN SOURCE

Coqui XTTS: Open Source Text-to-Speech with Voice Cloning

By

–

11 October 2023 18h50

18 – CoquiTTS (
@coqui_ai
) – @_josh_meyer_ and the Coqui team presented XTTS: A fully open source text to speech and voice generation model competitive with the commercial offerings, including instant-voice cloning http://
hf.co/spaces/coqui/X
TTS
…

Zama AI: Fully Homomorphic Encryption for Secure AI

By

–

11 October 2023 18h50

13 – Zama AI, presented by Benoit Chevallier-Mames – a company building open source fully homomorphic encryption (FHE) solutions for AI https://
zama.ai Specifically, their demo of a sentiment analysis application on encrypted data:

DINOv2: State-of-the-art open source self-supervised vision transformer

By

–

11 October 2023 18h50

9 – DINOv2 – presented by @TimDarcet
, DinoV2 is state of the art open source self supervised vision transformer model https://
dinov2.metademolab.com

Building Active Learning Pipelines with DagsHub and Hugging Face

By

–

11 October 2023 18h50

6 – @TheRealDagsHub : presented by @barazida and team, they showcase how one can build an Active Learning pipeline in the DagsHub platform, with open-source tools and then deploy the model to Hugging Face https://
dagshub.com

AcroBERT Model Disambiguates Acronyms in Conversations

By

–

11 October 2023 18h50

4 – Lihu Chen showcased AcroBERT , a model that disambiguates acronyms. If you have that friend that tries to sound smart by throwing a bunch of acronyms in the conversation, you can check out the demo and break them down https://
huggingface.co/spaces/Lihuche
n/AcroBERT
…

→ View original post on X — @clementdelangue

Mistral AI Paper Generates Nearly 500 Community Implementations

By

@clementdelangue

–

11 October 2023 18h38

Paper of the week by @MistralAI – almost 500 spaces based on it! You can ask questions about it to the authors here: https://
huggingface.co/papers/2310.06
825
…

→ View original post on X — @baiduresearch

HelixFold-Single: Fast MSA-Free Protein Structure Prediction

By

@baiduresearch

–

11 October 2023 18h00

Sharing our latest paper on fast protein structure prediction in @NatMachIntell
! We introduce HelixFold-Single, an open-source, MSA-free method that competes well with MSA-based methods but is significantly faster. Read the paper below.