You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Thanks for the paper and your job.
I have a question about the DNN-LSTM structure. You said in the paper that you stacked LSTM design will capture the temporal information and outperform DNN-only setting. But what I see from the code is that you shuffle the data in the provider file, which means the temporal information is lost. Also I didn't see you feed time sequence data into the network. It makes me confused. So I want to know how your cnn_lstm_block works exactly.
Hope for reply. Thanks.
The text was updated successfully, but these errors were encountered:
Thanks for the interest in our work! Yes, you should disable the shuffling line when you use the DNN-LSTM structure.
Note that we only release the code for DNN-only in this repo (see README.md). The cnn_lstm_block is just for reference. It is neither tested nor cleaned in this repo.
Thanks for the paper and your job.
I have a question about the DNN-LSTM structure. You said in the paper that you stacked LSTM design will capture the temporal information and outperform DNN-only setting. But what I see from the code is that you shuffle the data in the provider file, which means the temporal information is lost. Also I didn't see you feed time sequence data into the network. It makes me confused. So I want to know how your cnn_lstm_block works exactly.
Hope for reply. Thanks.
The text was updated successfully, but these errors were encountered: