Error in a multiclass classification in during of a prediction (Pycaret Predict function)

Question

I am a beginner in ML.

I have a small dataset with shape (200 (target values(5 classes), 100 (features) It's like a surface points in a part of a forest. They have geograpical coordianates. Also I have a features from 1000 other points. I need to predict target in this points.

I have done a basic steps from tutorial:

1. Set classification task with preprocessing steps
2. compare_models (n_select=3)
3. tuned_top3
4. bagged_top
5. blend
6. And finaly automl for finding a best model, then evaluate and finalize

After I try to predict target in 1000 point, but I get error.

How to solve this error? What I need to do for prepare an unseen data? They are not part of the Step 1 (Set function) I didn't find a solution in Documentaion for Pycaret. Thanks in advance.