[Docs] A question of Mini-InternVL

[xuliu-cyber]

📚 The doc issue

Hello, I read the paper "Mini-InternVL: A Flexible-Transfer Pocket Multimodal Model with 5% Parameters and 90% Performance", I wonder that during the domain adaptation phase, do you fine-tune three models separately in each domain? Thank you!

Suggest a potential alternative/fix

No response

[G-z-w]

We fine-tune models separately for each domain.