wangye8899 / bmec Goto Github PK

This repo contains BMEC dataset and related model code.

code -----> model and train code

data ------> BMEC dataset

results ------> experiments results

Loss curve

Accuracy curve

Firstly, you should create the same env according to the environment.yml

conda env create -f environment.yml

Secondly, you should unzip the BMEC dataset in data folder

cd data
unzip red_4_cells_raw.zip

Thirdly, you may revise the data path in train.py

parser.add_argument('--data_dir', metavar='DIR', default="your data path",
                    help='path to dataset')

Finally, you just run train.py to finish the training process.

nohup python train.py  > ./Log/Rotate/swin_transformer.log 2>&1 &

Our Shape Attention module is implemented in shape_attention.py

Classification Network is implemented in classification_net.py