how can we change the tokenizer to mclip so we have generate with Russian/Chinese/Japanese input?
· Sign up or log in to comment
