Saving multiple trained models with sklearn

Question

I'm trying to save multiple trained models for binary classification using pickle. I'm saving them in a pickle file as a tuple with the name of the model (name, model). However, when I load them and try to use the models all the predictions are zero. This is not the case if I train and predict all at once. Here is the part of the code where I'm saving the models:

    else:
        # Simple pipeline
        if dimension_reduction is None:
            model = Pipeline(steps=[('ss', scaler), ('clf', clf)])
        else:
            model = Pipeline(steps=[('ss', scaler), ('dr', dimension_reduction), ('clf', clf)])

        # Training the model  
        model.fit(X_train, y_train) 
        
        # Saving into a list
        models.append((name,model))

# Saving into a pickle file
with open("sklearn_models_{}.pckl".format(models_name), "wb") as f:
    for model in models:
        pickle.dump(model, f)
print('Models saved') 

I load them using

models = []
with open("sklearn_models_{}.pckl".format(NAME_MODELS), "rb") as f:
    while True:
        try:
            models.append(pickle.load(f))
        except EOFError:
            break 

The complete code is in this link. The training bit I put here is lines 166 to 183 and the loading bit lines 361 to 368.

It's there a better way to do this? What am I doing wrong?