AudioSeal is accepted at #ICML2024 ! 🚀 AI at Meta

AudioSeal is the state of art audio watermarking model designed for deepfakes mitigation. It is robust, lightning-fast & imperceptible⚡️

📄 paper: arxiv.org/abs/2401.17264

🔗 code (commercial friendly) github.com/facebookresear…

😍

Heat equation vs wave equation inside a planar domain. Boundary conditions matter for PDEs. en.wikipedia.org/wiki/Boundary_…

Just released MetaVoice-1B, the first text-to-speech model in candle. It can be used on CPU, Cuda GPUs, or Metal for macs, all that in pure #Rust ! It sounds a bit wobbly compared to the python version but still pretty nice!

Le Monde nous a rendu visite dans nos bureaux de l'hôtel d'Halwyll dans le Marais, pour en savoir plus sur notre mission.

Transformers outperform RNNs as they operate differently.
Do they?

Excited to share our new paper: “Transformers are Multi-State RNNs”

Paper: arxiv.org/abs/2401.06104
Code: github.com/schwartz-lab-N…

1/n

Very proud of the work done by Geoffrey Cideron , one of my last projects at Google. When we released MusicLM in May ’23, we incorporated a feedback system to realize the first ever large-scale, organic improvement of music generation through RLHF. 🎶🧵

We released a 70B version of CodeLlama today! Trained on 1T tokens, it is a much stronger base model for coding tasks. I look forward to seeing what the community will do with it! :)

Check out our latest work in music generation: MAGNeT 🧲!

MAGNeT is a single non-autoregressive transformer-based model for text-to-music and text-to-audio. Its' quality is on par with SOTA models, while it exceeds their speed by 7x.
pages.cs.huji.ac.il/adiyoss-lab/MA…

[1/6]

🧲 MAGNeT: Masked Audio Generation using a Single Non-Autoregressive Transformer 🎧 🤯 🎶 Colab 🥳

Thanks to Alon Ziv @ ICLR ❤ Itai Gat ❤ Gael Le Lan ❤ Tal Remez ❤ Felix Kreuk ❤ Alexandre Défossez ❤ Jade Copet ❤ Gabriel Synnaeve ❤ Yossi Adi ❤

🌐page: github.com/facebookresear…
📄paper:…

2023: highlights from our Brain team AI at Meta in a quick thread:

Want to convert from mono to stereo?
🔊➡️🔊🔊
In our latest work, we posit that upmixing mono to stereo is a great avenue for generative modeling, and that parametric stereo coding can facilitate things.

Paper: arxiv.org/abs/2306.14647

You can make great music, whether you're a shower singer or a charting artist. No instrument needed, just imagination.

Make your song today at suno.ai 🎧

As this paper is getting traction, let me share 5 major technical concerns in this thread:
openreview.net/pdf?id=WaLI8sl…