Tuwhy/Octopus-8B
Image-Text-to-Text
•
9B
•
Updated
•
11
RL checkpoints of Octopus-8B and baselines of paper: Learning Self-Correction in Vision–Language Models via Rollout Augmentation