Expertium

Here: https://drive.google.com/file/d/12yF5Lv-AbPqOnj9rp-_99TQHVxz7hqiU/view?usp=sharing It's a bit of a mess because I made some changes. The important files are other.py (where the `FSRS6` and `FSRS6ParameterClipper` classes are specified) and pretrain.py. In pretrain.py,...

> In fact, I have benchmarked two models with fractional time. You haven't published the results yet, right?

@fasiha we need a short 1-2 sentences summary of Ebisu, and a link to the repo or something like [this](https://fasiha.github.io/ebisu/) to use as a reference to your work.

@L-M-Sherlock @fasiha I benchmarked FSRS withh focal loss with different combinations of alpha and gamma. None of them improved RMSE or AUC.  Note that alpha=0.5 and gamma=0 is equivalent...

> I wonder if there's some kind of bug? Here's my implementation: ``` class FocalLoss(nn.Module): def __init__(self): super(FocalLoss, self).__init__() def forward(self, predictions: Tensor, labels: Tensor) -> Tensor: # hyper-parameters alpha...

I get an error when I try to run this: ``` from torchvision.ops import sigmoid_focal_loss class FocalLoss(nn.Module): def __init__(self): super(FocalLoss, self).__init__() def forward(self, predictions: Tensor, labels: Tensor) -> Tensor: #...

@fasiha in case you missed my comment here: https://github.com/open-spaced-repetition/srs-benchmark/issues/124#issuecomment-2367901327 > we need a short 1-2 sentences summary of Ebisu, and a link to the repo or something like [this](https://fasiha.github.io/ebisu/) to...

1) Given two cards of different users but with the same ID, should they be considered the same card? That is very unlikely to occur. IDs are generate from UNIX...

It's not exactly "trained on all 10k collections": we don't combine them into one giant collection. Instead, we optimize FSRS on every collection individually, and then take the median of...

@L-M-Sherlock there are some questions here that you should be able to answer better than me