search

GaiZhenbiao / Phi3V-Finetuning

Parameter-efficient finetuning script for Phi-3-vision, the strong multimodal language model by Microsoft.
Apache License 2.0
32 stars 8 forks source link

Multiple images? #10

Open pbarker opened 2 weeks ago

pbarker commented 2 weeks ago

Can this work for multiple images?

2U1 commented 2 weeks ago

@pbarker The official repo erased the example for multiple image, so I think it's not preferred for it. However, I'm working on it. I'll let you know, when I update the feature for it.

You could visit my repo for the update logs.

https://github.com/2U1/Phi3-Vision-ft

2U1 commented 2 weeks ago

@pbarker I've updated the code to support multi-image training and inference. You can use the above repo to train multi-image dataset.