resume from pretrained

[dodler]

Hi, is there an option to resume training from checkpoint?

[Starkguava]

use checkpoint?
with tf.Session(config=config) as sess:
checkpoint = tf.train.latest_checkpoint(params['checkpoint'])
saver.restore(sess, checkpoint)