Image Classification rps
Image Classification on the rock paper scissors dataset of tensorflow
My first try at image classification using tensorflow and keras.
Learned a lot about by convolution neural networks and model optimization.
My model accuracy:-
Model | Training Accuracy | Validation Accuracy | Additional Comments |
---|---|---|---|
1 | 96 | 77 | Basic CNN model |
2 | 99.8 | 84 | Increased number of layers |
3 | 99.8 | 81.2 | Integrated reduced LR |
4 | 100 | 82.2 | Integrated Early Stopping |
5 | 98 | 93.01 | Increased padding and MaxPooling from (2,2) to (3,3) |
7 | 98 | 93.28 | Increased the number of convolution layers and epochs |