uobinxiao

just wondering how to train the pre-train model from scratch. Does this repo contain pretraining code?

why only parts of the image_embeds are used?

1

https://github.com/Vision-CAIR/MiniGPT-4/blob/a431672c89f9c8fe4a6829c8c37f75724bb323aa/minigpt4/models/minigpt4.py#L139 As shown in the code, is there any reason that only parts of image_embeds are used?

Experimental results on ISIC dataset

3

Hi, as mentioned in the paper, the experimental results on ISIC dataset are listed in the appendix. But there is no appendix section in the priprint paper, any results of...

hello, as mentioned in the paper, this dataset follow the evaluation protocol of mediaeval 2020: Automatic polyp segmentation. However, their test set doesn't contain ground truth. And the dataset we...

how to speed up the evaluation

1

Thank you very much for sharing this useful benchmark. Currently, I am using this benchmark to evaluate our foundation model. However, I’ve noticed that both GPU and CPU utilization remain...

Hi, firstly, thank you so much for sharing the code. I am trying to reproduce the results reported in the paper, and came across some difficulties. From my understanding, the...