fault_diagnosis_cnn_pytorch

$$ step 1
raw singal >> time-frequence image

$$ step 2 
cropped time-frequence image getting training,valdiaton and test image

$$ step 3 
split train,val and test data; generate txt file
python data_pro/split_train_val_test.py

$$ step 4
train and validation
python train_cnn.py

$$ step 5
test
python predict.py