ValueError: bad marshal data

[YanZhangADS]

I am using Ubuntu 20.04 Linux

conda create -n eynollah -y python=3.8
conda activate eynollah
// in eynollah's directory
pip install -e .
make models

// execute following command generates "ValueError: bad marshal data"
eynollah -i data/input/1.jpg -o data/output -m models_eynollah

Full error messages is below.

Traceback (most recent call last):
File "/anaconda/envs/eynollah/bin/eynollah", line 33, in
sys.exit(load_entry_point('eynollah', 'console_scripts', 'eynollah')())
File "/anaconda/envs/eynollah/lib/python3.8/site-packages/click/core.py", line 1130, in call
return self.main(*args, **kwargs)
File "/anaconda/envs/eynollah/lib/python3.8/site-packages/click/core.py", line 1055, in main
rv = self.invoke(ctx)
File "/anaconda/envs/eynollah/lib/python3.8/site-packages/click/core.py", line 1404, in invoke
return ctx.invoke(self.callback, **ctx.params)
File "/anaconda/envs/eynollah/lib/python3.8/site-packages/click/core.py", line 760, in invoke
return __callback(*args, **kwargs)
File "/home/mylogin/notebooks/eynollah/qurator/eynollah/cli.py", line 151, in main
pcgts = eynollah.run()
File "/home/mylogin/notebooks/eynollah/qurator/eynollah/eynollah.py", line 2307, in run
img_res, is_image_enhanced, num_col_classifier, num_column_is_classified = self.run_enhancement()
File "/home/mylogin/notebooks/eynollah/qurator/eynollah/eynollah.py", line 1990, in run_enhancement
is_image_enhanced, img_org, img_res, num_col_classifier, num_column_is_classified, img_bin = self.resize_and_enhance_image_with_column_classifier()
File "/home/mylogin/notebooks/eynollah/qurator/eynollah/eynollah.py", line 408, in resize_and_enhance_image_with_column_classifier
_, page_coord = self.early_page_for_num_of_column_classification(img_bin)
File "/home/mylogin/notebooks/eynollah/qurator/eynollah/eynollah.py", line 648, in early_page_for_num_of_column_classification
model_page, session_page = self.start_new_session_and_model(self.model_page_dir)
File "/home/mylogin/notebooks/eynollah/qurator/eynollah/eynollah.py", line 518, in start_new_session_and_model
model = load_model(model_dir, compile=False)
File "/anaconda/envs/eynollah/lib/python3.8/site-packages/keras/utils/traceback_utils.py", line 70, in error_handler
raise e.with_traceback(filtered_tb) from None
File "/anaconda/envs/eynollah/lib/python3.8/site-packages/keras/utils/generic_utils.py", line 103, in func_load
code = marshal.loads(raw_code)
ValueError: bad marshal data (unknown type code)

[bertsky]

I am also hitting this. Looks like a similar problem we countered with sbb_binarization. But it's strange this occurs on 3.8 already.

Anyway, I propose we switch to SavedModel format quickly. Did anyone already publish something?