I'm working with a previously saved model, however, when I try to load the model, I receive the following error message:
---------------------------------------------------------------------------
ImportError Traceback (most recent call last)
<ipython-input-11-f8d353bad309> in <module>
----> 1 estimator, corpus_encoder, label_encoder, tokenizer = load_model(MODEL_PATH, loss=weighted_binary_crossentropy)
<ipython-input-6-03f944f0a51f> in load_model(model_base_dir, loss, optimizer, metrics)
9 json_file.close()
10 model = model_from_json(loaded_model_json)
---> 11 model.load_weights(os.path.join(MODEL_PATH, 'model.h5'))
12 model.compile(loss=loss, optimizer=optimizer, metrics=metrics)
13 # loading context_encoder
~\anaconda3\envs\davis_env\lib\site-packages\keras\engine\network.py in load_weights(self, filepath, by_name, skip_mismatch, reshape)
1154 """
1155 if h5py is None:
-> 1156 raise ImportError('`load_weights` requires h5py.')
1157 with h5py.File(filepath, mode='r') as f:
1158 if 'layer_names' not in f.attrs and 'model_weights' in f:
ImportError: `load_weights` requires h5py.
I have since install cython and h5py as described in the answers found here and here
Are there any other solutions?
I'm using tensorflow version 1.12.0 and keras version 2.2.4
UPDATE:
When simply trying to import h5py (via ```import h5py``) I receive the following error message:
ImportError Traceback (most recent call last)
<ipython-input-15-c9f0b8c65221> in <module>
----> 1 import h5py
~\anaconda3\envs\davis_env\lib\site-packages\h5py\__init__.py in <module>
32 raise
33
---> 34 from . import version
35
36 if version.hdf5_version_tuple != version.hdf5_built_version_tuple:
~\anaconda3\envs\davis_env\lib\site-packages\h5py\version.py in <module>
15
16 from collections import namedtuple
---> 17 from . import h5 as _h5
18 import sys
19 import numpy
h5py\h5.pyx in init h5py.h5()
ImportError: DLL load failed: The specified procedure could not be found.```