Thanks so much @chris_j_paxton and @micoolcho for inviting me on the latest @RoboPapers episode. I had a fantastic time talking through VaViM & VaVAM. I really enjoyed our lively, deep-dive conversation and the opportunity to showcase what we’re building @valeoai 🚗🤖✨!.

RT @Ar_Douillard: I'll discuss distributed learning on Saturday, July 12. First, I'll cover current methods needing high bandwidth, then n….

RT @ycombinator: François Chollet (@fchollet) on the ARC Prize and how we get to AGI. At AI Startup School in San Francisco. 00:00 - The….

RT @leloykun: This effect seems to just be an artifact of SGD/Adam/AdamW/etc and more modern optimizers, e.g. Muon/Shampoo/PSGD, don't have….

RT @brianchristian: Reward models (RMs) are the moral compass of LLMs – but no one has x-rayed them at scale. We just ran the first exhaust….

RT @torchcompiled: Modeling dolphin language is cool. Translating it into human speak is cooler. Somewhere you're gonna want to figure ou….

RT @RobertTLange: Text-to-LoRA: What if you no longer had to fine-tune your LLM for every single downstream task?. 🚀 Stoked to share our wo….

RT @xunhuang1995: Real-time video generation is finally real — without sacrificing quality. Introducing Self-Forcing, a new paradigm for t….

RT @jxmnop: new paper from our work at Meta!. **GPT-style language models memorize 3.6 bits per param**. we compute capacity by measuring t….

RT @seohong_park: We found a way to do RL *only* with BC policies. The idea is simple:. 1. Train a BC policy π(a|s).2. Train a conditional….

RT @edward_milsom: To address the "parameterisation lottery" (ideas win because they work well with popular choices of e.g. learning rates….

RT @OpenDriveLab: 💥 Forget slow autoregression and skip rigid full-sequence denoising! Nexus is a next-gen predictive pipeline for realisti….

RT @ylecun: A talk on Self-Supervised Learning.

RT @rgilman33: The secret life of SwiGLU . Simple neurons like those using ReLU, GELU or SiLU create a new dimension, then slice across th….

RT @micoolcho: This makes the past 3 years of work and often months away from my family worth it. A big shoutout to @NoriakiHirose , Lydi….

RT @RoboPapers: Full episode dropping soon!. Geeking out with @zhiyuan_zhou_ on AutoEval: Autonomous Evaluation of Generalist Robot Manipul….

RT @francislee2020: Introducing #UniVLA, this is a unified VLA framework that enables policy learning across different environments, exhibi….

RT @IlirAliu_: A robot hand grasp over 500 totally new objects without fail? Zero-shot, single-view & super reliable .⬇️ + Paper. Grasping….

RT @arankomatsuzaki: Google just released a new eval for video generation called TRAJAN. TRAJAN does automated evaluation of temporal cons….

RT @rgilman33: Do diffusion models want Perlin-like noise?. SDXL-turbo spends the first blocks simply creating Perlin-like noise of differe….