CODEWITHSUNDEEP
neural-networktensorflowlstmrecurrent-neural-network
spiralarchitect
spiralarchitect

Reputation: 910

Tensorflow translate training - when will it stop?

I am using example from Google's Tensorflow. I am able to run it well, but the training doesn't seem to stop as it has a while True: loop. reference: translate.py

Coe in train():

while True:
  # Choose a bucket according to data distribution. We pick a random number
  # in [0, 1] and use the corresponding interval in train_buckets_scale.
  random_number_01 = np.random.random_sample()
  bucket_id = min([i for i in xrange(len(train_buckets_scale))
                   if train_buckets_scale[i] > random_number_01])
....

Upvotes: 0

Views: 809

Answers (1)

Max
Max

Reputation: 560

This is true, at one point you can abort the training process. The moment of it depends on the model parameters you have choosen. According to the tensorflow documentation it takes about 340K steps with choosen batch-size of 64. If have selected the option to create checkpoints, you can use any of the checkpoints to check your models performance with the decoder.

Upvotes: 1

Related Questions