output_path = "../logs/afri-rlhf"
if not os.path.exists(output_path):
os.makedirs(output_path)
dataset.save_to_disk(output_path)
It seems train, test and validation splits of the datasets are being saved the same directory, each time a dataset saved it overwrites the previous dataset. we may need to save them in different directory.
It seems train, test and validation splits of the datasets are being saved the same directory, each time a dataset saved it overwrites the previous dataset. we may need to save them in different directory.