Twitxy
Saved
@MacPawResearch's profile picture
View on X

MacPaw AI & Research

Verified account This account is verified. Learn more

@MacPawResearch

The research department at MacPaw.
On-device AI, LLM efficiency, AI memory, HCI.
Ukraine research.macpaw.com Joined October 2015
72 Following 181 Followers 77 Posts

Latest posts 5

MacPaw AI & Research Verified account @MacPawResearch
Happy to see our research in action 😉
Binoui @Binouilol Oct 22
@MacPawResearch Been experimenting a lot with genRA and loving it! It really has enabled me to create my own animations. I now use a sheet of key poses generated as a reference for the key poses generation. Then I decide if those keys are good enough for motion afterwards
Video preview
View quoted post on X
MacPaw AI & Research Verified account @MacPawResearch
It was awesome presenting two of our papers today at #ECCV

- GenRA: End-to-end Framework for LLM-Driven Animation of Rigged 3D Models - research.macpaw.com/publications/g…
- Rigid 3D Object Alignment: Optimization vs. Feed-Forward Prediction - research.macpaw.com/publications/3…

So many new ideas on applications in robotics, realistic 3d scene generation, sign language and of course usage with GPT-6 Astra ✨

If you want to catch us live - see you tomorrow on Wild3D poster session at 14:50!
2 replies 2 reposts 18 likes 161K views View on X
MacPaw AI & Research Verified account @MacPawResearch
Your Mac can run a 26B diffusion LLM now📷

We just added DiffusionGemma 26B-A4B to Fast-dLLM-mlx, our open-source MLX inference framework for diffusion LLMs on Apple Silicon. Inference is ~30% faster than mlx-vlm 0.6.4 on M4 Pro, and the speedup is training-free (dual cache + parallel decoding with probability thresholding).

Charts and repro commands: github.com/MacPaw/Fast-dL…