#LSTM MODEL
7 messages · Page 1 of 1 (latest)
LSTM is a neural network which is mostly used to process sequential data
unlike regular neural networks which each neuron's output only depends on it's current input, the output of each neuron in LSTM depends both on it's current input and also the outputs of the previous neurons
so it has a memory which can remember the outputs of the previous neurons
so for example given an incomplete sentence and asks the LSTM to predict the next word
let's say "He went to the store to buy ..."
each word in the sentence contributes to predicting the next word