Issue Uploading Custom Model with Pipeline on AI VERIFY

[macotochou]

Hi everyone,

I'm encountering an issue while trying to upload a simple custom MLP model (Tensorflow Keras Sequential) with a pipeline on AI VERIFY. Despite following the tutorial to save the pipeline (using 'train.py'), I receive an error notification stating "There was an error deserializing the model" when uploading the 'custom_model_pipeline' file.

Interestingly, the process works fine when I execute 'test.py' on my PC, and the model successfully predicts the testing data without any deserialization problems.

Here is the structure of my project for reference:
https://drive.google.com/file/d/1KqwS9Q6kdwwyGJ83G5aFh1kuR1vEY6cF/view?usp=drive_link
discussion/
│
├── custom_model_pipeline/
│ ├── binary_classification_pipeline.sav
│ ├── CustomClass.py
│ └── # The file I upload to AI VERIFY
│
├── train.py
│ └── # Script for saving binary_classification_pipeline.sav
│
├── test.py
│ └── # Script for loading binary_classification_pipeline.sav
│
└── data/
├── binary_classification_pipeline_training.sav
├── binary_classification_pipeline_ytrain.sav
├── binary_classification_pipeline_testing.sav
└── binary_classification_pipeline_ytest.sav

Here are the versions of the software I'm using:
Python:3.11.9
TensorFlow:2.12.0
Scikit-learn:1.2.2

Has anyone else encountered this issue or have any suggestions on how to resolve it? Any help would be greatly appreciated!

Thanks in advance.