Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention Paper • 2502.11089 • Published 28 days ago • 143
Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention Paper • 2502.11089 • Published 28 days ago • 143
TripoSG: High-Fidelity 3D Shape Synthesis using Large-Scale Rectified Flow Models Paper • 2502.06608 • Published Feb 10 • 33
Running on Zero 434 434 Chat with DeepSeek-VL2-small 🌍 Generate responses using images and text input