GRADEO: Towards Human-Like Evaluation for Text-to-Video Generation via Multi-Step Reasoning

Open in new window