maneprajakta / Image_Classification

All about Image classification problem 's solved during TensorFlow learning.

Image_Classification

All about image classification problems solved using TensorFlow while learning TensorFlow. -Dataset of every problem is included in folder

1.Happy_or_sad_emoji:

-Binary Classification
-A cnn with use of ImageDataGeneratorClass of tensorflow.
-Source dataset contains sad and happy dataset 40 images each train on them and used for prediction.

2.Cat_or_Dog :

-Binary Classification
-A cnn with intermediate representation of layers of model.
-Trying to get better validation accuracy using agumentation.
-Source data set conatains train and validation images seprate and each contain cat and dog seprate images . data : https://www.kaggle.com/c/dogs-vs-cats/
-Prediction

3.Fashion_Mnist :

-Multiclass Classification.
-First without cnn than using cnn to get better accuracy .
-Using callback to stop training when desired acc reached.
-fashion_mnist from keras dataset.