- Built several CNN models to predict Image-Recognition
- Data processing to transform training and testing dataset.
- Python 3.8
- Packages: pandas, numpy, seaborn, matplotlib, keras
- KerasAPI
- multi class loss
- early stop
- adjust parameter
- Youtube
- Divided train and test data set by 255 to fit the CNN model
- Zoom in, rotate, flip, and changed the shear range with the train data set, then fit the testing dataset.