Sandro Cavallari

@shaohua0116 @chelseabfinn @jessezhang @robertarail Any available recording ?

RL is hitting a ceiling with human feedback. What if the world itself becomes the signal? Introducing RLxF: RL from World Feedback 🌍 workshop at ICML 2026 in Seoul, South Korea! Speakers: David Silver, @chelseabfinn, @jessezhang, @robertarail Web page: sites.google.com/view/rlxf-icml…

@svlevine @berkeley_ai Would be great if you could share a recording link

Tomorrow in the Workshop on World Models at ICLR in Rio (10:30 am) I’ll talk about a… different take on what might make for a good world model. Come find out, 10:30 in Room 202A at ICLR sites.google.com/view/iclr-2026…

Building with FLUX.2: From Prototype to Real-World Application x.com/i/broadcasts/1…

Introducing FLUX.2 [klein]. Blazing fast. Beautiful. Generate stunning images in under a second while maintaining exceptional quality. Great for fast editing, changing styles, and developing ideas from 0 → 1. Available via API, or run it locally - Klein 4B under Apache 2.0, Klein 9B as open weights. Try it for free in our demo app (link in the thread).

Incredible to see 90 GB pipeline running on a 5080. Big congratulation to @bfl_ml for such an amazing model and for partnering with us and @ComfyUI to pull this off !!! blogs.nvidia.com/blog/rtx-ai-ga… #NVIDIA #NVIDIAstory

FLUX.2 just dropped and we've teamed up with @bfl_ml + @ComfyUI to optimize performance on RTX AI PCs. 🖼️🔥 Get started today in the #RTXAIGarage blog: nvda.ws/49AIeOk

Join in on @bfl_ml’s FLUX.1 Kontext [dev] Hackathon! 🙌 Create an incredible local use case for a chance to win a NVIDIA GeForce RTX 5090 GPU. 🎉🏆

Great job by our engagers in speeding up comfyUI inference remaining in pytorch native

📢 @bfl_ml's FLUX.1 Kontext image editing model just dropped as an NVIDIA NIM. Grab it in @ComfyUI and edit any image using a few words or even another image. 📚 Get started in our #RTXAIGarage blog: nvda.ws/4luzhZg

📸 Generate. Refine. Repeat. @bfl_ml’s FLUX.1 Kontext simplifies AI image editing with text + image prompts—now RTX-accelerated with TensorRT. Bonus: Google’s new Gemma 3n multimodal LLM is also accelerated on RTX. 🔗 #RTXAIGarage: nvda.ws/4lCEM8M

High quality image editing no longer needs closed models We release FLUX.1 Kontext [dev] - an open weights model for proprietary-level image editing performance. Runs on consumer chips. ✓ Open weights available ✓ Best in-class performance ✓ Self-serve commercial licensing

Today we're releasing FLUX.1 Kontext - a suite of generative flow matching models that allow you to generate and edit images. Unlike traditional text-to-image models, Kontext understands both text AND images as input, enabling true in-context generation and editing.

The amount of advertisements on YouTube is insane !!!

Really fun work on FP4 quantization and Blackwell ... and more is coming developer.nvidia.com/blog/nvidia-te…

🚀Join us for NVIDIA #GTCParis at Viva Technology 2025—Europe’s biggest startup and tech event. June 10-12, immerse yourself in visionary sessions and hands-on demos showcasing the latest AI innovations. Explore the impact of AI across Europe and beyond. bit.ly/4mlnE8t

TensorRT-LLM delivers fast, flexible LLM inference - but the full pipeline can be complex. This dstack example simplifies it end-to-end: build the container, convert the model, and deploy on any cloud or on-prem. Showcases both DeepSeek R1 and its distilled Llama version. dstack.ai/examples/deplo…

🚀 New drop: @bfl_ml FLUX.1-dev is now available as a NVIDIA NIM microservice — turning your RTX AI PC / Workstation into an AI image-gen powerhouse. 🎬 Plus: New RTX-accelerated AI features in the DaVinci Resolve Studio 20 beta. 🔗 #RTXAIGarage blog: nvda.ws/3EgCwnu

Can someone explain to me why people pay for online courses made by unqualified people about topic such as: - get rich with Bitcoin - how to invest in X - crate passive income with ... (name any AI tech) when top universities/professors provides recorded classes for free ?

@y0b1byte @huggingface Is zig any good to program in cuda ?

I can load @huggingface safetensors natively in zig now!

📣Just announced at #CES2025: #NVIDIACosmos Build custom world models for physical #AI systems at scale. Get started today. bit.ly/3DHjru1