flexitok--mod-tokenizers-individual_overlap.json

{"9": {"ratio_to_total_tokens": 1.0, "expected_training_ratio_in_superset": 1.0, "num_tokens": 17}, "total_training_compared_to_full_model": 1.0}