
🇫🇷 ML Imagineer • Making ML demos actually usable • Hugging Face Fellow @huggingface • Master Animation / VFX / Motion @psl_univ @ecole_artsdeco
Paris
Joined November 2009
- Tweets3.8K
- Following1K
- Followers7.7K
- Likes13.7K
Sylvain Filoni retweeted
Papers with Code is now open-source!
Apache 2.0-license.
Let's collaboratively make the best paper interface!
github.com/huggingface/paper…
Directing an actor shouldn’t require decoding an interface. 🎙️
Let’s try TAKE STUDIO: a clearer flow from script to performance. Write, direct, listen, refine.
Fish Audio’s Drama 3 · @Gradio · ZeroGPU @huggingface
—> huggingface.co/spaces/fffilo…
Show this @huggingface Space to your mom.
Thank me later. 🤗
—> huggingface.co/spaces/fffilo…
For the HF Fellows 🤓
Qwen3.5-9B: photo analysis & fragrance concept
d3-lite: optional coherence checks
PUP: FLUX prompt refinement
FLUX 3 Image Klein / Z-Image-Turbo: bottle images
Custom @Gradio UI + ZeroGPU.
Shown here: Z-Image-Turbo.
I baked a Carbon-X Playground 🧬
5 experiments to explore DNA with Hugging Face Bio’s Carbon-X. Try curated samples or your own sequences, ask questions and compare answers.
Built with @Gradio, running on ZeroGPU @huggingface Space 🤗
—> huggingface.co/spaces/fffilo…
Introducing Carbon-X: The best open DNA model.
We midtrained a 14B model on DNA and text, and it beats every open language model and matches GPT-6.1 Sol at 1,000x lower cost.
Ask it about a sequence, what a mutation does, or to generate new DNA.
huggingface.co/HuggingFaceBi…
Sylvain Filoni retweeted
Jensen spent 10 min of his keynote asking microducks to sit, then explained Nvidia's roadmap to them 🥹
Sylvain Filoni retweeted
Cool dataset release from Scenario: 1,342 high quality AI-generated 3D assets on Hugging Face (CC BY 4.0).
we need more of this let's build huge shared libraries of 3D objects so we don't waste tokens generating the same things over and over :)
Sylvain Filoni retweeted
the amount of economic value created by open-source AI is unmatched
Fun and crazy napkin math: Meta, NVIDIA, Alibaba, DeepSeek, Moonshot, Mistral and ~30 other labs have shared close to $1B worth of training compute on the Hugging Face Hub
Free to download
End of 2023 it was under $40M
Wikipedia was written in volunteer hours. This AI age commons is written in GPU-hours and data
(it's a floor btw: about half comes from official compute numbers rest is educated guess and it still miss the ~900k community fine-tunes)
Space with data to explore: thomwolf-hub-dollars.static.…
Nice, but you know the rule:
If it’s not on @huggingface shelves,
it does not exist.
💡
Introducing SKY-001, the first Engineering World Model.
On CAD and BIM tasks, it beats GPT-6 Astra and every frontier model we tested:
-> 51-125x faster
- > 150-210x cheaper
-> Higher IOU than all frontier models
SKY-001 lets you dream a physical object before you build it.
18 months of work. Built by the team out of Maluuba's deep learning lab (acq. by @MSRResearch).
The first ‘AI General Engineer’ will be the backbone for Autonomous Manufacturing.
I just gave you 4 @huggingface Banger @gradio Spaces to play with and enjoy this weekend. 🙏
Grok said: Really smart way to showcase Decision Models 👌
Using Kev-4B to score free-form player actions (no text generation) while Ink handles the story structure and Qwen only does the dialogue is exactly the right hybrid approach. Fun, clear, and fully local. Nice work @fffiloni 🎲🤗



















