Training instruction for IRSAM

We organize the training folder as follows.

train
|____datasets
|____segment_anything_training
|____demo.py
|____utils
| |____dataloader.py
| |____log.py
| |____misc.py
| |____metric.py
| |____metrics.py
| |____loss_mask.py
| |____misc.py
|____workdirs
|____mobile_sam.pt

According to the given train/test.txt, place the training/testing images/labels into four separate folders.

The current IRSAM_encoder provided is the baseline of Mobile-SAM, and our designed IRSAM_encoder with WPMD will be released soon.

Evaluation

To evaluate IRSAM on various dataset, modify the dataset path in the train-IRSAM.py file

Example

python train_IRSAM.py --output workdirs/your_workdir --checkpoint your_checkpoint --eval