CreatePredictionEngine Fails due to error "One of the identified items was in an invalid format"

**System Information (please complete the following information):**
- Model Builder Version: Tried both 16.14.4.2312403 on VS 2019 and the latest on VS 2022
- Visual Studio Version (if applicable): Tried on VS 2022 and VS 2019

**Describe the bug**
- On which page of the tutorial did you run into an issue (URL): https://dotnet.microsoft.com/en-us/learn/ml-dotnet/get-started-tutorial/consume
- Clear description of the problem:
-I was trying to consume the trained model in my project. The auto-generated console app was fine. I built the model with my project into a dll. When I was using the dll to do the prediction, I kept meeting the issue: "One of the identified items was in an invalid format".
It happens at "CreatePredictionEngine".

**To Reproduce**
Steps to reproduce the behavior:
1. After training and evaluating the model, add the code into my own project which targets at net 2.0.
2. Build the project.
3. Reference the built dll of the project in 2, and call the prediction method
4. See error:

**Expected behavior**
Give prediction results.

**Screenshots**
If applicable, add screenshots to help explain your problem.


**Additional context**
Add any other context about the problem here.

Inizia con il consume tutorial e riproduci il problema in CreatePredictionEngine usando l’applicazione console generata, la DLL del modello compilato e un progetto net 2.0. Confronta il framework di destinazione e gli assembly referenziati tra l’applicazione generata funzionante e il progetto che non funziona. Il lavoro è completato quando la DLL referenziata può creare un motore di predizione e restituire risultati di predizione senza l’errore di formato non valido.

Scritto dal modello di indicizzazione a partire dal testo della issue.