TensorFlow2.0_SSD icon indicating copy to clipboard operation
TensorFlow2.0_SSD copied to clipboard

After one epoch, there is a problem, and can't save the model.

Open jiollos opened this issue 4 years ago • 3 comments

Traceback (most recent call last): File "train.py", line 86, in ssd.save_weights(filepath=save_model_dir+"epoch-{}".format(epoch), save_format="tf") File "C:\Users\jeo13\Anaconda3\lib\site-packages\tensorflow\python\keras\engine\network.py", line 1167, in save_weights self._trackable_saver.save(filepath, session=session) File "C:\Users\jeo13\Anaconda3\lib\site-packages\tensorflow\python\training\tracking\util.py", line 1187, in save file_prefix=file_prefix_tensor, object_graph_tensor=object_graph_tensor) File "C:\Users\jeo13\Anaconda3\lib\site-packages\tensorflow\python\training\tracking\util.py", line 1127, in _save_cached_when_graph_building object_graph_tensor=object_graph_tensor) File "C:\Users\jeo13\Anaconda3\lib\site-packages\tensorflow\python\training\tracking\util.py", line 1095, in _gather_saveables feed_additions) = self._graph_view.serialize_object_graph() File "C:\Users\jeo13\Anaconda3\lib\site-packages\tensorflow\python\training\tracking\graph_view.py", line 381, in serialize_object_graph trackable_objects, path_to_root) File "C:\Users\jeo13\Anaconda3\lib\site-packages\tensorflow\python\training\tracking\graph_view.py", line 337, in _serialize_gathered_objects object_names[obj] = _object_prefix_from_path(path) File "C:\Users\jeo13\Anaconda3\lib\site-packages\tensorflow\python\training\tracking\graph_view.py", line 64, in _object_prefix_from_path for trackable in path_to_root)) File "C:\Users\jeo13\Anaconda3\lib\site-packages\tensorflow\python\training\tracking\graph_view.py", line 64, in for trackable in path_to_root)) File "C:\Users\jeo13\Anaconda3\lib\site-packages\tensorflow\python\training\tracking\graph_view.py", line 57, in _escape_local_name return (name.replace(_ESCAPE_CHAR, _ESCAPE_CHAR + _ESCAPE_CHAR) AttributeError: 'NoneType' object has no attribute 'replace'

jiollos avatar Jul 19 '20 06:07 jiollos

我也是同样的问题

sunxianpeng123 avatar Aug 10 '20 09:08 sunxianpeng123

that's my first comment on github so: you need to put name of weights in order to save model (keras bug) self.add_weight(shape=(1, 1, 1, 512), dtype=tf.float32, initializer=tf.keras.initializers.Ones(), trainable=True, name='fckr')

after this you can save it

raduciou avatar Oct 31 '20 15:10 raduciou

the name in the file of ssd.py(line numer:13),like this: "self.learnable_factor = self.add_weight(shape=(1, 1, 1, 512), dtype=tf.float32, initializer=tf.keras.initializers.Ones(), trainable=True,name='ssd_train')"

xiaotaonan avatar Dec 11 '20 09:12 xiaotaonan