Can MobileVLM v2 receive multiple images at once？

[wuwu-C]

No description provided.

[wuwu-C]

And can I give it history conversation to acheive in-context inference

[wyddmw]

MobileVLM/mobilevlm/model/mobilevlm.py

Line 100 in 688fdec

if type(images) is list or images.ndim == 5:

According to this implementation, MobileVLM can receive multiple images at once in default, but you need to modify the dataloader to load multiple input images in a list or introduce an additional dimension.