tokenizer_config.json

{"vocab_size": 50000, "min_freq": 2, "special_tokens": ["<pad>", "<eos>", "<bos>", "<unk>"]}