When I first ran the H2OAutoML framework, that is, using the default AMLB installation, an error appeared indicating that OpenML datasets could not be read in ARFF format. It seems the issue is that Java, which H2O is based on, cannot access the datasets if the import_file function is used in the exec.py file. If this is replaced with upload_file, it works correctly. I’m not sure if it’s the best solution, but it is a solution.
When I first ran the H2OAutoML framework, that is, using the default AMLB installation, an error appeared indicating that OpenML datasets could not be read in ARFF format. It seems the issue is that Java, which H2O is based on, cannot access the datasets if the import_file function is used in the exec.py file. If this is replaced with upload_file, it works correctly. I’m not sure if it’s the best solution, but it is a solution.