Readme file

[xuqianchj]

Hello, I am a beginner and I am very interested in your project. I want to use the pre-trained weights you provided to predict an ultrasound image, but the readme file you provided is too simple and I cannot achieve prediction on the image.

[18066085680]

Hello, I am a beginner and I am very interested in your project. I want to use the pre-trained weights you provided to predict an ultrasound image, but the readme file you provided is too simple and I cannot achieve prediction on the image.

I am also meet this problem ,it's quiet difficult for me to run this project , it's excellent if Owner public a details file

[nvblanco]

I agree with the previous comments. It would be nice to have information in this regard, at least to use the pretrained endoViT module as a deep feature extractor.

[Chop1]

Hi,

for downstream, we have configs files in the repo. You can just modify the class and data path according to your need.

https://github.com/CAMMA-public/UltraSam/blob/main/configs/UltraSAM/UltraSAM_full/downstream/classification/BUSBRA/UltraSam.py

For more detail about the config system, you can visit the mmEngine doc.

I will update the readme with a quickstart / template guide for common use cases, once the paper is accepted.

[xuqianchj]

I tried what you said, and then ran the mim test mmdet configs/UltraSAM/UltraSAM_full/UltraSAM_point_refine.py --checkpoint ./work_dirs/UltraSam/iter_30000.pth code. I changed the weight file to the pre-trained weight file you provided, and an error occurred.