tokenizer_config.json

{"do_lower_case": false, "vocab_type": "gpt2", "model_max_length": 512}