CODEWITHSUNDEEP
pythontensorflowreturnn
Patrick von Platen
Patrick von Platen

Reputation: 82

How to load the weights of a trained network in Returnn with TensorFlow

When trying to load the saved weights over multiple epochs of a trained networks in returnn using the following code: 

import tensorflow as tf
from returnn.Config import Config
from returnn.TFNetwork import TFNetwork

for i in range(1,11):
    modelFilePath = path/to/model/ + 'network.' + '%03d' % (i,)

    returnnConfig = Config()
    returnnConfig.load_file(path/to/configFile)
    returnnTfNetwork = TFNetwork(config=path/to/configFile, train_flag=False, eval_flag=True)

    returnnTfNetwork.construct_from_dict(returnnConfig.typed_value('network'))

    with tf.Session() as sess:
        returnnTfNetwork.load_params_from_file(modelFilePath, sess)

I get the following error: 

Variables to restore which are not in checkpoint:
global_step_1

Variables in checkpoint which are not needed for restore:
global_step

Probably we can restore these:
(None)

Error, some entry is missing in the checkpoint

Upvotes: 1

Views: 157

Answers (1)

Albert
Albert

Reputation: 68110

The problem is that you recreate TFNetwork every time in the loop, and there, also a new variable is created every time for the global step, which must be called different because every variable must have a unique name.

You could do something like this inside the loop:

tf.reset_default_graph()

Upvotes: 1

Related Questions