Expertium
Expertium
Ooops, sorry, I referenced the wrong issue. Don't pay attention to the "mentioned this issue" thing above.
> converting Anki ratings (1-4) into something Ebisu handles I'm not great at math, but how about just assigning different values of q0 for each button? In other words, each...
``` def halflife2stability(h): return h / 24 / np.log(0.5) * np.log(0.9) ``` I don't think it works this way. In Ebisu, you need alpha and beta to convert stabilities/halflives, and...
Regarding grade mapping, I have suggested using different values of q0 for each grade.
I think Sherlock means that as long as delta_t and S are exactly the same, the probability of recall is also exactly the same. If a card with delta_t=x (last...
@L-M-Sherlock @fasiha I'm just reminding you in case you were busy and forgot about this issue.
@fasiha me and LMSherlock are reworking RMSE and I just wanted to ask whether you still want to participate in the benchmark.
Well, I'm not the code expert here, that's LMSherlock.
I have an idea how to make FSRS more flexible, but I'm currently testing other ideas too, so I'll probably start working on Hard/Easy next week.
This isn't the idea that I mentioned in my previous comment, this is another idea. I wonder what you all think https://github.com/open-spaced-repetition/fsrs4anki/issues/461#issuecomment-1858933062