v1.3.0

Updates to codec_bpe.audio_to_codes tool:

• Support batched inference via --batch_size argument
• Automatically attempt to infer the --codec_type argument value from the --codec_model argument
• Output codec_info.json file to simplify usage of downstream tools such as codec_bpe.train_tokenizer