Just wanted to say thanks for the model!I also noticed that even though it was fine-tuned on a Japanese dataset, it works surprisingly well on English text.Maybe the dataset still contained a few English samples?
· Sign up or log in to comment
