Skip to content

Latest commit

 

History

History
20 lines (15 loc) · 812 Bytes

README.md

File metadata and controls

20 lines (15 loc) · 812 Bytes

Convolution_LSTM_pytorch

Thanks for your attention. I haven't got time to maintain this repo for a long time. I recommend this repo which provides an excellent implementation.

Usage

A multi-layer convolution LSTM module Pytorch implementation of Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting

clstm = ConvLSTM(input_channels=512, hidden_channels=[128, 64, 64], kernel_size=5, step=9, effective_step=[2, 4, 8])
lstm_outputs = clstm(cnn_features)
hidden_states = lstm_outputs[0]

Thanks

Thanks to @Jackie-Chou and @chencodeX who provide lots of valuable advice. I apology for the inconvenience.