Creating a Text Classifier With LSTM! PyTorch Deep Learning Tutorial

[@waliqadri]

Amazing video very nice.. thanks for putting so much effort.

[@LukeDitria]

Thank you!

[@FranzOberhauser12]

Hello, I really like the video. My question is that why did you assign hidden and memory to variables if you are not going to use them? I have a lstm model too and I am having a problem with it so I was wondering id it is due to not using hidden and memory from the output.

[@LukeDitria]

I'm not sure what you mean? We do use the hidden and memory, they get passed into the LSTM.

[@FranzOberhauser12]

@@LukeDitria I might be mistaken, but after you return 'hidden' and 'memory' from the model, when the batch is completed and the loop restarts, aren't both redefined to zeroes? If we go over the variable 'hidden,' for example, get it from the model during predictions, it is not used, and when the loop is reset, it resets to zeroes.

[@LukeDitria]

Well yes, these are returned by the LSTM as the hidden and memory buffer from that final step. I am just getting the model to return them here, but no they are not later used. This is just a leftover from previous code examples where I do use them, but it can be ignored for this case.

[@FranzOberhauser12]

@@LukeDitria I see, I got confused since my model is not working and I was curious if It was related to it. Can I ask about my problem on your dc?

[@LukeDitria]

@@FranzOberhauser12 yes that would be the best place to discuss.

[@テムル-c3w]

thanks for the video.

[@ChideraObiezue]

Nice video, pls more on this

[@LukeDitria]

Thanks for the feedback, is there anything in particular you'd like to see?