2 Layer LSTM

[Tobias128128]

In the article "LSTM-based Encoder-Decoder for Multi-sensor Anomaly Detection" from Malhotra are two LSTM layers (encoder and decoder) mentioned. You refered to that article and you wanted to implement that concept. But I can't see in your RNNPredictor-model that you implement two separate LSTM layer. Do you deviate from that concept intentionally or have I missed something?

[chickenbestlover]

Sorry to confuse you. I used only the part to calculate the anomaly score of the paper. 2018년 12월 31일 (월) 오전 1:36, Tobias128128 <notifications@github.com>님이 작성:

…

In the article "LSTM-based Encoder-Decoder for Multi-sensor Anomaly Detection" from Malhotra are two LSTM layers (encoder and decoder) mentioned. You refered to that article and you wanted to implement that concept. But I can't see in your RNNPredictor-model that you implement two separate LSTM layer. Do you deviate from that concept intentionally or have I missed something? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#20>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AJ2X1DmbCmM-vBPOOc1nWHpeUn2zqII4ks5u-Ot2gaJpZM4Zk_sg> .

[Tobias128128]

Thank you for your response.
Do you think that the architecture in this paper from Malhotra is more powerful?

[chickenbestlover]

You mean enc-dec model is better than rnn prediction model? Yes, I assume that it is because 'reconstruction' is easier than 'prediction'. 2018년 12월 31일 (월) 오후 10:37, Tobias128128 <notifications@github.com>님이 작성:

…

Thank you for your response. Do you think that the architecture in this paper from Malhotra is more powerful? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#20 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AJ2X1EWnIhnfuxieeLWeag42gEjHvVHFks5u-hMdgaJpZM4Zk_sg> .

[Tobias128128]

I‘ve implemented the lstm autoencoder model in pytorch and it seems to work great with machine data. I‘m now validating the model...

[rockytallen]

I‘ve implemented the lstm autoencoder model in pytorch and it seems to work great with machine data. I‘m now validating the model...

I have read the article 'LSTM-based Encoder-Decoder for Multi-sensor Anomaly Detection'.
And interested in implementing the model to test also.
could you please share with me the information(code) in detail.
many thanks

[dagap]

I‘ve implemented the lstm autoencoder model in pytorch and it seems to work great with machine data. I‘m now validating the model...

Did you use a Sequence2Sequence like model for this?