Damn, seriously, AI is out of control, I've never seen such perfect lip-sync with emotion in AI 😮 Meta just dropped : MoCha: Towards Movie-Grade Talking Character Synthesis - First ever: Multi-character, turn-based dialogue - Talking characters 8 wild examples & details 👇

Thanks a lot for sharing our work! Appreciate! The original post is here x.com/CongWei1230/st…

Meta just announced MoCha This AI can create full movie-quality talking & singing characters from just speech & text. 10 wild examples: 1. Talking Characters

it's over.. actors this new AI by Meta can generate talking actors using text or voice and.. you can control their emotion & acting, make them sing and even generate 2 character dialogue 10 unreal examples:

ThursdAI - Apr 3 - OPEN... AI? Amazon Nova-act, MCP update from WB, Runway Gen 4, Nomic Embed & more AI news x.com/i/broadcasts/1…

Thanks a lot for sharing our work!

Thanks for sharing our work!

🎬MoCha generates Movie-grade talking Characters **directly from Text and Speech** (please unmute) 🌟MoCha supports **multi-character turn-taking**, like in movies. Project page: congwei1230.github.io/MoCha/

Had a lot of fun to scale up RL to improve math reasoning! Excited to introduce AceMath-RL-Nemotron-7B with a scalable training recipe 📑Full blog: research.nvidia.com/labs/adlr/acem… 🔗Model: huggingface.co/nvidia/AceMath…

Veo 3 blew people's mind in generating talking characters! It's so exciting! But we need evaluation benchmark for that! Cong just released the MochaBench used in our Mocha paper to evaluate talking character models. Github: github.com/congwei1230/Mo… Mocha Paper:

I made a tool called Diffusion Explorer that lets you to train and visualize simple 2D diffusion and flow models live in the browser. You can draw your own distributions and observe how the generated samples converge during training. Try it live 👇

with just math-RL, AceReason-Nemotron-14B surpass DeepCoder-14B on LiveCodeBench v5. we then did code-RL and found training becomes so much easier

Introducing QuickVideo, 🚀 speeding up the end-to-end time from the mp4 bit stream to VideoLLM inference by at least 2.5 times for hour-long video understanding (e.g 1024 frames) on a single 40GB GPU. 🤔What are the key challenges of hour-long video understanding? 1.

Introducing VerlTool - a unified and easy-to-extend tool agent training framework based on verl. Recently, there's been a growing trend toward training tool agents with reinforcement learning algorithms like GRPO and PPO. Representative works include SearchR1, ToRL, ReTool, and

🤔 Feel like your AI is bullshitting you? It’s not just you. 🚨 We quantified machine bullshit 💩 Turns out, aligning LLMs to be "helpful" via human feedback actually teaches them to bullshit—and Chain-of-Thought reasoning just makes it worse! 🔥 Time to rethink AI alignment.