LSTM by Example using Tensorflow
Rowel Atienza
71814

Looks like wrong inspiration by Aymeric Damien?? Why are you taking just the softmax from the last output node as predicition?

You should do prediction at each output node. The error calculation will be different then as you have to take the summation of deltas from each timestep into account.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.