- Input shape: (32, 32, 3)
- Label shape: (10)
- Input shape: (32, 32, 3)
- Label shape: (10)
Layer | Output Shape | Connected to |
---|---|---|
input_1 | (?, 32, 32, 3) |
Layer | Output Shape | Connected to |
---|---|---|
conv2d_1 | (?, 32, 32, 16) | input_1 |
batch_normalization_1 | (?, 32, 32, 16) | conv2d_1 |
activation_1 | (?, 32, 32, 16) | batch_normalization_1 |
conv2d_2 | (?, 32, 32, 16) | activation_1 |
batch_normalization_2 | (?, 32, 32, 16) | conv2d_2 |
activation_2 | (?, 32, 32, 16) | batch_normalization_2 |
conv2d_3 | (?, 32, 32, 16) | activation_2 |
batch_normalization_3 | (?, 32, 32, 16) | conv2d_3 |
add_1 | (?, 32, 32, 16) | activation_1, batch_normalization_3 |
activation_3 | (?, 32, 32, 16) | add_1 |
max_pool_1 | (?, 16, 16, 16) | activation_3 |
Layer | Output Shape | Connected to |
---|---|---|
conv2d_4 | (?, 16, 16, 32) | max_pool_1 |
batch_normalization_4 | (?, 16, 16, 32) | conv2d_4 |
activation_4 | (?, 16, 16, 32) | batch_normalization_4 |
conv2d_5 | (?, 16, 16, 32) | activation_4 |
batch_normalization_5 | (?, 16, 16, 32) | conv2d_5 |
activation_5 | (?, 16, 16, 32) | batch_normalization_5 |
conv2d_6 | (?, 16, 16, 32) | activation_5 |
batch_normalization_6 | (?, 16, 16, 32) | conv2d_6 |
add_2 | (?, 16, 16, 32) | activation_4, batch_normalization_6 |
activation_6 | (?, 16, 16, 32) | add_2 |
max_pool_2 | (?, 8, 8, 32) | activation_6 |
Layer | Output Shape | Connected to |
---|---|---|
conv2d_7 | (?, 8, 8, 64) | max_pool_2 |
batch_normalization_7 | (?, 8, 8, 64) | conv2d_7 |
activation_7 | (?, 8, 8, 64) | batch_normalization_7 |
conv2d_8 | (?, 8, 8, 64) | activation_7 |
batch_normalization_8 | (?, 8, 8, 64) | conv2d_8 |
activation_8 | (?, 8, 8, 64) | batch_normalization_8 |
conv2d_9 | (?, 8, 8, 64) | activation_8 |
batch_normalization_9 | (?, 8, 8, 64) | conv2d_9 |
add_3 | (?, 8, 8, 64) | activation_7, batch_normalization_9 |
activation_9 | (?, 8, 8, 64) | add_3 |
avg_pool_1 | (?, 1, 1, 64) | activation_9 |
Layer | Output Shape | Connected to |
---|---|---|
flatten_1 | (?, 64) | avg_pool_1 |
dense_1 | (?, 10) | flatten_1 |
softmax_1 | (?, 10) | dense_1 |
- Average train cost: 0.073 (at 30 epoch)
- Train accuracy: 0.9800 (at 30 epoch)
- Test accuracy: 0.8037 (at 30 epoch)