Training cubist and randomforest models produces 'temporary' pickle files that are required to make predictions with these models. These files are placed in a directory called 'results' in the current working directory. Predictions are then hardcoded to look for this directory - to predict with a cubist or RF model you require not only the '.model' but also this 'results' directory to be in the expected place.
These files should be included as part of the model itself.
Training cubist and randomforest models produces 'temporary' pickle files that are required to make predictions with these models. These files are placed in a directory called 'results' in the current working directory. Predictions are then hardcoded to look for this directory - to predict with a cubist or RF model you require not only the '.model' but also this 'results' directory to be in the expected place.
These files should be included as part of the model itself.