-
Notifications
You must be signed in to change notification settings - Fork 277
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
get_replacement_mapper_ in seq_aligner.py #53
Comments
I also have same question. Can anyone answer for us? :) |
I think this is because your source prompt and target prompt are almost the same, and the processed results of the tokenizer are so similar to the space-splited ones. For these cases, it will return the diagonal matrix. However, If the source text is "a lion is eating an apple", and the edited text is "a lovely-dog is eating an apple", the "lion" will be mapped to "lovely", "-", and "dog", because the tokenizer will tokenized the "lovely-dog" into "lovely", "-", "dog". |
I have the same question... |
you are right bro! |
I have the same question,too. |
|
Thanks for this amazing work. I tested many times on this func, and this func always returns a diagonal matrix with 1s on the diagonal. Why don't you use the built-in func in the torch directly if this is right? If this needs to be corrected, can you help explain this issue?
The text was updated successfully, but these errors were encountered: