Cat-vs-Dog-Classification

use CNN / DNN to classify the label of the image (cat/dog); image argumentation / transfer learning -> 95% accuracy or above

Configuration

• tensorflow == 2.7 & python == 3.8.0 (very important as tensorflow > 2.7 changes preprocess api)
• pillow (any version) for image loader
• matplotlib, numpy for data processing and graph

Setup

Manually

• Download the cats-vs-dogs zip from https://storage.googleapis.com/mledu-datasets/cats_and_dogs.zip
• Manually delete two dbs in ImagePet/Dogs and ImagePet/Cats

Linux

• !wget --no-check-certificate
  "https://download.microsoft.com/download/3/E/1/3E1C3F21-ECDB-4869-8368-6DEBA77B919F/kagglecatsanddogs_5340.zip", notebook will automatically download it.
• !find /tmp/PetImages/ -type f ! -name "*.jpg" -exec rm {} +' two dbs in the pets image will automatically delete

Transfer learning

• Download the model weight data from https://storage.googleapis.com/mledu-datasets/inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5

Tips

• Don't recommand run it locally!!!
• Otherwise run it applying GPUs:

pip install tensorflow-gpu
import tensorflow as tf
print("Num GPUs Available: ", len(tf.config.experimental.list_physical_devices('GPU')))