Our first RNN!
Create inputs
len(xs), xs[0].shape
(8, (75110,))
but at:
Our first RNN with keras!
model.fit(np.concatenate(xs,axis=1), y, batch_size=64, nb_epoch=8)
(i was surprised that we can do “np.concatenate(xs,axis=1)”.)
So I did add xs[0].shape again, it becomes:
(75110, 1)
When did this happen???