LSTM Model Initialization fails if hidden_dims are anything except [1,1,1]

[iSiddharth20]

When Initializing LSTM Model in main.py :
model_lstm = ConvLSTM(input_dim=1, hidden_dims=[1,1,1], kernel_size=(3, 3), num_layers=3, alpha=0.5)
The model doesn't work if hidden_dims are anything except [1,1,1] when num_layers=3

Possible Source of Error :

• __init__ of Class ConvLSTMCell of lstm_model.py
• init_hidden and forward of Class ConvLSTM of lstm_model.py