2 Different Models to Identify Shape
layer Structure
Model 1 - 5 Layer Structure
layer 1 - Conv2D - 32 filters, 3x3 kernel, ReLU
layer 2 - MaxPooling2D - (2x2)
layer 3 - Flatten
layer 4 - Dense - 64 units, ReLU
layer 5 - Dense - 3 units, softmax
Model 2 - 8 Layer Structure
layer 1 - Conv2D - 32 filters, (3x3) kernel, ReLU
layer 2 - MaxPooling2D - (2x2)
layer 3 - Conv2D - 64 filters, (3x3) kernel, ReLU
layer 4 - MaxPooling2D - (2x2)
layer 5 - Flatten
layer 6 - Dense - 128 units, ReLU
layer 7 - Dropout - rate (0.3)
layer 8 - Dense - 3 units, softmax
Training
Model 1
Accuracy: 0%
Loss: 0%
epochs: 0
training images: 0
Training not started.
Model 2
Accuracy: 0%
Loss: 0%
epochs: 0
training images: 0
Training not started.
Train Model 1
Results
Model 1 Results
Model 2 Results
Test Model
Circle, square or triangle.
Predict
Clear
Please wait for the model to train, then wait again to load weights into its neural net—good things come to those who wait!.