Most diffusion research today asks: How can we sample faster? But I think another question is equally important: Are we training diffusion models in the right way? famous-bubbler-dcc.notion.si…

Today we're releasing ZAYA1-74B-Preview, a major milestone in scaling pretraining on @AMD. ZAYA1-74B-Preview is a 4B active / 74B total MoE. This preview model is a strong pre-RL base checkpoint. The final post-trained reasoning model is coming soon. 🧵

Really excited to see Uni-1 out in the world 🔥Our first unified model. The range of things this model can do is wild: image-to-~100 styles, manga generation, multi-ref with strong identity preservation, temporal storytelling, sketch-to-image, spatial reasoning, multilingual infographics, layering… the capability range is honestly unreal. this is just the start 🫡 check out the blog to learn more lumalabs.ai/uni-1 Proud of the team and what we’re building at @LumaLabsAI 🚀

🤯💥

I’m currently in transit to San Diego for NeurIPS. If you’re also killing time, feel free to check out a 2-minute-30-second horror sci-fi short film Michael and I recently created. We’d love any comments or likes: devpost.com/software/dreamca… Looking forward to catching up at the venue! 🎥

Why must robots be human-shaped? Bringing impossible creatures into the real world can create just as beautiful an emotional bond ❤️

I feel the debate shouldn’t only be about whether DiT is effective, but also about how information preservation is the key to accelerating diffusion training. Our MicroDiT (arxiv.org/abs/2407.15811) paper showed this: by letting masked token info mix into unmasked ones, we can cut down a lot of tokens with only minor performance loss. Interestingly, two months ago, when I caught up with @StefanABaumann at #CVPR, we discussed how TREAD and MicroDiT are conceptually similar from info perspective. Maybe it’s time to look at diffusion through an information-theoretic lens: from post-training (for the better alignment) to latent space curation, I believe this could lead to some really exciting discoveries!

Shout out for Doji!

Excited to introduce Reka Vision, an agentic visual understanding and search platform. Transform your unstructured multimodal data into insights and actions.

Introducing our V1 Video Model. It's fun, easy, and beautiful. Available at 10$/month, it's the first video model for *everyone* and it's available now.

@cveu_workshop starting at 1:00 PM, 207 A-D.

Heading to Nashville 🎸 for @CVPR (06/11 - 06/16)! Always excited to catch up with old friends and make new connections. Let’s grab a coffee ☕️ or chat about diffusion models, post-training, or just life! #CVPR2025 #Diffusion #GenerativeAI #Nashville

you're now closer to the year 2050 than the year 2000

📢 Our paper "Training Ultra Long Context Language Model with Fully Pipelined Distributed Transformer" has been accepted to hashtag#MLSys2025, taking place May 12-15! Excited to share our research at the intersection of machine learning and systems in San Jose, CA. 🎉 Check out the full program here: lnkd.in/ecyzGbwJ hashtag#MLSys hashtag#MachineLearning hashtag#Systems hashtag#Conference

Delighted to see MicroDiffusion paper being accepted at CVPR. Checkout the code and models if you are looking for an extremely low cost setup for latent diffusion models.

Our **Flow Matching Tutorial** from #NeurIPS2024 is now publicly available: neurips.cc/virtual/2024/tuto… @helibenhamu @RickyTQChen

Following fully open-source philosophy, we’ve released the official training code, data code, and model ckpts for our micro-budget training of diffusion models from scratch (MicroDiTs). Now anyone can train a Stable Diffusion v1/v2-quality model from scratch in just 2.5 days using 8 H100 GPUs (<$2000 cost). Github: github.com/SonyResearch/micr… Checkpoints: huggingface.co/VSehwag24/Mic… @SonyAI_global 1/3