Cannot parse keras metadata models

[julfontaine]

Hi there,

I'm running into an error when attempting to execute the main script with the following command:
python main.py -xgb -odds=fanduel

However, I receive the following error related to loading a saved Keras model:

OSError: Cannot parse keras metadata at path Models/NN_Models/Trained-Model-ML-1699315388.285516\keras_metadata.pb: Received error: Error parsing message

I've tried with different model and I always have the error parsing message.

I'm using Python 3.11 on a Windows system. Tensorflow and Keras are both version 2.14.0.
Has anyone else faced this issue, or could the maintainers please advise on how to resolve this?

Any help would be greatly appreciated!
Related issue - 4 days ago

error log:

Traceback (most recent call last):
  File "D:\NBA-Machine-Learning-Sports-Betting\main.py", line 9, in <module>
    from src.Predict import NN_Runner, XGBoost_Runner
  File "D:\NBA-Machine-Learning-Sports-Betting\src\Predict\NN_Runner.py", line 12, in <module>
    model = load_model('Models/NN_Models/Trained-Model-ML-1699315388.285516')
  File "\AppData\Local\Programs\Python\Python311\Lib\site-packages\keras\src\saving\saving_api.py", line 262, in load_model
    return legacy_sm_saving_lib.load_model(
  File "\AppData\Local\Programs\Python\Python311\Lib\site-packages\keras\src\utils\traceback_utils.py", line 70, in error_handler
    raise e.with_traceback(filtered_tb) from None
  File "\AppData\Local\Programs\Python\Python311\Lib\site-packages\keras\src\saving\legacy\saved_model\load.py", line 113, in load
    raise IOError(
OSError: Cannot parse keras metadata at path Models/NN_Models/Trained-Model-ML-1699315388.285516\keras_metadata.pb: Received error: Error parsing message

[umut1tr]

I have the same Problem on my Machine. I could not really dig into yet :(

[rpassaro]

Sorry this is late, just saw it. Run it through WSL on windows if you can. You can look up how to do this or ask chatgpt but this fixed it for me!

[umut1tr]

@rpassaro thanks for the hint!

[egamxmx]

hello,

try the follow steps:

1. path: \NBA-Machine-Learning-Sports-Betting

cd src/Process-Data
python -m Get_Data
python -m Get_Odds_Data
python -m Create_Games

cd ../Train-Models
python -m XGBoost_Model_ML
python -m XGBoost_Model_UO

3. cd NBA-Machine-Learning-Sports-Betting\Models\NN_Models

delete all folder, just leave Trained-Model-ML, Trained-Model-OU

4. cd src/Train-Models

python -m NN_Model_ML.py
python -m NN_Model_UO.py

5. update NN_Runner.py with the correct path in the line 12 and 13

model = load_model('Models/Trained-Model-ML-')
ou_model = load_model("Models/Trained-Model-OU-")

6. run: python main.py -nn -odds=fanduel

[yampara]

Still throws same error with new path:

Traceback (most recent call last):
File "/mnt/d/BettingApp/NBA-Machine-Learning-Sports-Betting/main.py", line 9, in
from src.Predict import NN_Runner, XGBoost_Runner
File "/mnt/d/BettingApp/NBA-Machine-Learning-Sports-Betting/src/Predict/NN_Runner.py", line 12, in
model = load_model('Models/NN_Models/Trained-Model-ML')
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/lsaizalv/.local/lib/python3.11/site-packages/keras/src/saving/saving_api.py", line 262, in load_model
return legacy_sm_saving_lib.load_model(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/lsaizalv/.local/lib/python3.11/site-packages/keras/src/utils/traceback_utils.py", line 70, in error_handler
raise e.with_traceback(filtered_tb) from None
File "/home/lsaizalv/.local/lib/python3.11/site-packages/keras/src/saving/legacy/saved_model/load.py", line 113, in load
raise IOError(
OSError: Cannot parse keras metadata at path Models/NN_Models/Trained-Model-ML/keras_metadata.pb: Received error: Error parsing message
lsaizalv@DESKTOP-I8UKAIU:/mnt/d/BettingApp/NBA-Machine-Learning-Sports-Betting$

[yampara]

Found solution, use WSL with kali-linux distribution and download the project from zip file not git clone. Somehow git clone with windows throws that error.