Pavlo Molchanov

Latest

• Nemotron-Labs-Diffusion: A Tri-Mode Language Model Unifying Autoregressive, Diffusion, and Self-Speculation Decoding
• Recurrence without Recurrence: Stable Video Landmark Detection with Deep Equilibrium Models
• LANA: Latency Aware Network Acceleration
• A-ViT: Adaptive Tokens for Efficient Vision Transformer
• See through Gradients: Image Batch Recovery via GradInversion