Table 1 Parameter of STDConvLSTM model.
From: Deep spatio-temporal dependent convolutional LSTM network for traffic flow prediction
Hyperparameters | Value |
|---|---|
Number of filters in CNN/DCNN | 16,8/8,2 |
CNN layers/ConvLSTM layers/DCNN layers | 2 |
Stride | 1 |
Convolutional kernel size | (3,3) |
Hidden channels of ConvLSTM | 64 |
Padding | 1 |
Batch size | 32 |
Epoch (TaxiNYC) | 800 |
Epoch (BikeNYC) | 500 |
Optimizer | Adam |
Learning rate | 0.0001 |
