search

BAAI-DCAI / Bunny

A family of lightweight multimodal models.
Apache License 2.0
799 stars 61 forks source link

interlaced information between images and text and multiple images #79

Open zhangqingwu opened 1 month ago

zhangqingwu commented 1 month ago

I want to train a model that can understand interlaced information between images and text. The prompt words may contain multiple images. Can the current code be configured to train multiple images? If it needs to be modified, how to modify it is more reasonable.

LAW1223 commented 1 month ago

The current code version only supports single-image training. The multi-image training and evaluation capabilities will be included in a future release, and the code will be made publicly available at that time.

RussRobin commented 1 week ago

Hi @zhangqingwu , thank you for your interest in Bunny. In SpatialBot we release Bunny codes for multi-image. Please note that interlaced training is not included in it. Also, SpatialBot codes will not be kept up-to-date with Bunny in the future.

Regards Russell