ByteDance's Seedvr 2 7B is an open-source, one-step diffusion-based video restoration model that performs adversarial training against real data, leveraging an adaptive window attention mechanism to handle high-resolution video. It is genuinely best at achieving comparable or better performance than existing video restoration approaches in a single step.
Input
Output
Context
-
Max Output
-
Parameters
7B
Input Modalities
Output Modalities
Estimates based on INT8 quantization. Actual requirements vary by framework and configuration.
Data sourced from official provider APIs and documentation
Last updated: Jun 23, 2026
Automatically route workloads to the right model for every task, every time.