tokenizer_config.json

{"do_lower_case": true, "keep_accents": true, "model_max_length": 512}