when I specify asix=1 in cross_entropy of neural_network_raw it does not work
grvmeratwal opened this issue · comments
grvmeratwal commented
As you have used axis=1 in cross_entropy of logistic_regression.ipynb
If I do same thing in neural_network_raw , it does not work
it gives accuracy of 12%
return tf.reduce_mean(-tf.reduce_sum(y_true * tf.math.log(y_pred),axis=1))
grvmeratwal commented
return tf.reduce_mean(-tf.reduce_sum(y_true * tf.math.log(y_pred),axis=1))
does not works in neural_network_raw
grvmeratwal commented
when I make learning_rate=0.05
it works with axis=1