-
-
Notifications
You must be signed in to change notification settings - Fork 123
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
[BUG] [FSRS4Anki v1.6.1 Optimizer] Optimized Default Difficulty is too low on large revlog #31
Comments
But the default difficulty deceased monotonically in your case. I will test your collection tomorrow. |
Hm actually, you might be right. A large majority of my cards have ~250% ease And looking at the interval history if I keep pressing Good:
The change between each interval is:
This seems correct since most of my cards have 250% and I have about ~88% retention rate for my young cards. The FSRS4anki algorithm starts with 250% change for the interval and then it slowly decreases the change in interval as I keep pressing good. In version FSRS4Anki v1.4.3 Optimizer however, I got the results:
The change between each interval here is:
With these parameters, it starts with a change in interval of 200%, not 250%. It does seem like these new parameters are correct. I guess the larger learning rate allowed it to converge faster. |
Describe the bug
After optimizing the parameters using
fsrs4anki_optimizer
on a large revlog, the default difficulty is too lowTo Reproduce
Steps to reproduce the behavior:
Output:
Expected behavior
The default difficultly should not be as low as 1.1873
Environment
Additional context
I believe the issue has come from the commit 4185f3d
Change the learning rate from
lr=1e-4
tolr=5e-4
might be too aggressive?The text was updated successfully, but these errors were encountered: