ByteDance just released EvoQuality on Hugging Face
A self-evolving vision-language model for image quality assessment.
It learns from its own predictions, no human labels required,
using voting and GRPO to push past supervised baselines.
https://t.co/HGMdrlTIhz
Congrats to @GoogleDeepMind on the launch of DiffusionGemma.
The model generates 256 tokens in parallel per step, delivering 150+ TPS on DGX Spark, and 1,000+ TPS on a single H100.
We're supporting it from day one with:
• BF16 and NVFP4 checkpoints on @huggingface🤗
• Free GPU-accelerated endpoints on https://t.co/6T0R9P7EXS
• @vllm_project support with FP8 precision
Get started with DiffusionGemma on NVIDIA: https://t.co/vurk7GCQUs
Gemma goes diffusion! DiffusionGemma with up to 1000+ tokens per second! 🌬️
- Built on Gemma 4 as a 26B MoE model.
- 3.8B parameters during inference.
- Generates text in 256-token blocks in parallel.
- Fits within 18 GB VRAM limits when quantized.
- Apache 2.0
"R³: 3D Reconstruction via Relative Regression"
TL;DR: replaces global pose regression with confidence-weighted relative pose estimation, enabling scalable streaming and offline 3D reconstruction with only 372M parameters.
Nex-N2 is now open source!An agentic model series from Nex AGI built for coding, tool use, deep research, and long-horizon workflows. 🧠🔎
🛠️ https://t.co/p2n2cMdBlR
⚙️ https://t.co/k67uqVgDDP
● Models: Nex-N2-Pro 397B total, 17B active; Nex-N2-mini 35B total, 3B active
● Agentic Thinking: adaptive reasoning depth + coherent reasoning across coding, search, tool calling, and execution
● Efficiency: Nex-N2-mini saves roughly 20% overall token cost vs forced thinking while matching or slightly exceeding task performance
● Open-model lead: 75.3 on Terminal-Bench 2.1, 80.8 on SWE-Bench Verified, 83.7 on BrowseComp, and 1585 on GDPval among listed open baselines
● Deployment: customized SGLang fork, reasoning parser, tool-call parser, Docker image
● License: Apache 2.0
BREAKING:
Jensen Huang just told retail to buy SpaceX at IPO.
"Like buying Amazon. Google. Meta in the early days."
But here is what that comparison leaves out.
Amazon's earliest investors paid $0.09 per share.
Meta IPO'd at $38. Fell to $17 within months.
SpaceX is targeting a $2,000,000,000,000 valuation at IPO.
That is not the early days.
The early money already made its return in private.
Insiders own 95% of shares.
Lockup period: just 60 days.
By November 2026 most insider shares could already be sold.
Retail gets to buy at peak valuation.
While insiders get liquidity.
The early days comparison is real.
The early days pricing is not.
Do your own research before you buy.
I want to offer some unsolicited advice to computer vision researchers jumping into robotics. Don't focus too much on VLMs, VLAs etc. That's fine, but the real action is at the sensorimotor level. Most of the open problems in robotics are in manipulation, which is about hand-object interaction, and contacts and forces are central. Proprioception and tactile sensing are as important as vision. Don't get seduced by cherry-picked demos. You can't do robotics without doing robotics.
NVIDIA just released the Anchor Lab dataset on Hugging Face
Real-world robotics measurements to calibrate simulation against physical data for zero-shot sim-to-real deployment.
https://t.co/K32ETnbYKk