You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi authors, thank you for releasing this nice work. The paper's findings seemed to be based on the linear beta schedule, and I'm wondering if the method would also work with other beta schedules, like sigmoid or cosine?
The text was updated successfully, but these errors were encountered:
Thank you for your question, we try other beta schedules and report the results in section 3.3 cross-architecture robustness evaluation. Our method works on different beta schedule like 'linear', 'cosine', 'quadratic'.
If there are any other questions, feel free to contact us.
Hi authors, thank you for releasing this nice work. The paper's findings seemed to be based on the linear beta schedule, and I'm wondering if the method would also work with other beta schedules, like sigmoid or cosine?
The text was updated successfully, but these errors were encountered: