Blame for tokenizer_config.json

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