MultiCatClassifier

A simple model to classify different cat species using pretrained resnet models.

Introduction

We used the pre-trained ImageNet ResNet model, changed its fully connected layer, and got 95.90% accuracy on the test set.

Datasets

5 classes of cats.

Ragdolls	Singapura cats	Persian cats	Sphynx cats	Scottish fold cats

Train

Just run the train.py

python train.py

Remember to replace data_path with your own path. the default path is './cats'.

Test

Just run the test.py

python test.py

Here you need to define test_data, test_labels yourself. You can refer to the method in train.py to read in, or you can read in according to your own method.