How can i train further using the pretrained checkpoint files?

[hemanthadawale]

I tried to load the checkpoint files for further training:
In train.py >
try: global_step = tf.contrib.framework.load_variable("model", "model.ckpt-300000") except Exception as err: global_step = 0

Gets called to checkpoint_utils.py >
def _get_checkpoint_filename(filepattern, name): """Returns checkpoint filename given directory or specific filepattern.""" if gfile.IsDirectory(filepattern): return saver.latest_checkpoint(filepattern, latest_filename = name) return filepattern

Gets called to saver.py >
def get_checkpoint_state(.....): try: if file_io.file_exists(coord_checkpoint_filename): print('*****file_io.file_exists******') file_content = file_io.read_file_to_string(coord_checkpoint_filename)

The file is not being detected.
But on specifying the complete path in train.py : (adding '.meta' or '.index' or '.data-00000-of-00001') >
global_step = tf.contrib.framework.load_variable("model", "model.ckpt-300000.meta")

the file is detected, but the next line 'file_io.read_file_to_string(coord_checkpoint_filename)' throws > " 'utf-8' codec can't decode byte 0xc3 in position 1: invalid continuation byte "

Can anyone help me out with this issue?
Thank you

[hemanthadawale]

@leix28 please help if you know how to train further from checkpoints.