use transfer learning for kaggle cats_vs_dogs
first using ResNet50 to extract features from data and store in features.hdf5 file.Then create logistic regression classifier to classify cats and dogs
- keras
- scikit-learn
- h5py
- imutils
- numpy
- progressbar
first run extract_features.py
using following command
python extract_features.py --dataset datasets/train --output datasets/features.hdf5
it will create file name features.hdf5
inside hdf5 folder
then run train_model.py
using following command
python train_model.py --dataset datasets/features.hdf5 --model datasets/dogs_vs_cats.pickle