Dear authors,
Many thanks for your inspiring work. Our group is eager to follow your research, but we were unable to find the code for training the TA-Tok tokenizer and de-tokenizer. Could you kindly release that portion of the code?
Thank you very much for your time and consideration!