This repository contains the code for the FungiCLEF and SnakeCLEF competitions.
dataset is read from the *.csv file in "./metadata" directory, which includes the labels and storage paths of the images.
1 using the following command to train your own model
python train_seesawloss.py
2 Select the *.pth file from the "./checkpoint" directory and modify the value of the CFG dictionary in "predict.py"
python predict.py
dataset is read from the *.csv file in "./metadata" directory, which includes the labels and storage paths of the images.
1 using the following command to convert the meta text information to a vector
python extract_metadata_feature.py
1 using the following command to train your own model
python train_pyramid_meta.py
3 using the following command to train your own prior model
python train_prior.py
4 sellect *.pth file from "./checkpoint" directory and modify the value of the CFG['checkpoints'] and CFG['prior_checkpoint'] in "predict.py". Then, using the the following command to predict
python predict.py