TriAttention: Efficient Long Reasoning with Trigonometric KV Compression Paper • 2604.04921 • Published Apr 6 • 114
Tuna-2: Pixel Embeddings Beat Vision Encoders for Multimodal Understanding and Generation Paper • 2604.24763 • Published 28 days ago • 71
Seedance 2.0: Advancing Video Generation for World Complexity Paper • 2604.14148 • Published Apr 15 • 162
VecGlypher: Unified Vector Glyph Generation with Language Models Paper • 2602.21461 • Published Feb 25 • 12
VGGRPO: Towards World-Consistent Video Generation with 4D Latent Reward Paper • 2603.26599 • Published Mar 27 • 66
Running on CPU Upgrade 236 The Synthetic Data Playbook: Generating Trillions of the Finest Tokens 📝 236 Explore synthetic data experiments on a virtual bookshelf
Running on CPU Upgrade Featured 3.18k The Smol Training Playbook 📚 3.18k The secrets to building world-class LLMs
One-step Latent-free Image Generation with Pixel Mean Flows Paper • 2601.22158 • Published Jan 29 • 18
UniVideo: Unified Understanding, Generation, and Editing for Videos Paper • 2510.08377 • Published Oct 9, 2025 • 81
HiStream: Efficient High-Resolution Video Generation via Redundancy-Eliminated Streaming Paper • 2512.21338 • Published Dec 24, 2025 • 23