SAM-LST

Pytorch implementation of paper Ladder Fine-tuning approach for SAM integrating complementary network.

We will update the code progressively.

Data Preparation

We borrow the data process from SAMed.

Training & Testing

Training

python train.py --root_path <Your folder> --output <Your output path> --warmup --AdamW 

Testing

python test.py --is_savenii --output_dir <Your output directory> --my_ckpt <ckpt path>

Pretrained Models

• Pretrained Models are available at https://drive.google.com/file/d/1sjZhgp02j4Pp_Uvk5Bhg50bELafWu-sR/view?usp=drive_link.

Acknowledgements

This repo is based on SAMed and TransUnet.

Thanks original authors for their work!