roomylee / rnn-text-classification-tf

Tensorflow Implementation of Recurrent Neural Network (Vanilla, LSTM, GRU) for Text Classification
MIT License
118 stars 51 forks source link

Number of training steps #3

Open montaser85 opened 6 years ago

montaser85 commented 6 years ago

How can I define the number of training steps? In my attempt, the code has stopped running after 16000 steps and a final accuracy is shown. how can I increase or decrease these steps? accuracyvanilla

roomylee commented 6 years ago

@montaser85 total training steps are changed by batch_size and num_epochs These are calculated by the below equation. total steps = num_epochs * (size of dataset / batch_size)

https://github.com/roomylee/rnn-text-classification-tf/blob/35a8a48d163a75f2c3bc3bfd1c42e3999c866e95/train.py#L28