bidur-khanal / densenet-lstm Goto Github PK

`
sequential_image= Input(shape=(17,120,200,1))
sequential_coordinate= Input (shape= (17,4))

base_network = DenseNet(dense_blocks=5, dense_layers=-1, growth_rate=8, dropout_rate=0.2,
bottleneck=True, compression=1.0, weight_decay=1e-4, depth=40,batch_norm=False)

encoded_patches= TimeDistributed(base_network)([sequential_image,sequential_coordinate])

LSTM_layer= LSTM(256, return_sequences=True)(encoded_patches)
Dense1= Dense(128, activation ='linear')(LSTM_layer)
outputs= Dense (8, activation='linear')(Dense1)
model= Model(inputs= sequential_image,outputs= outputs)

return model
`
Could not wrap two-input network with TimeDistributed. sequential_image and sequential_coordinate have differnet size.
There is no issue if the wrapped network is a single input network
@bishesh