Nicolas Hug

Grid search classes should accept estimator instances instead of just classes

2

Scikit learn does, surprise doesn't, and that sucks :) Current workaround for changing default parameter values that you want to keep constant across searches, e.g. with random_state: `grid = {'random_state':...

Add adjusted cosine similarity

9

As discussed in #135, it would be interesting to add support for the adjusted cosine similarity. Adjusted cosine is similar to pearson similarity, only the centering differs (for a user-user...

For now only strings are accepted as the `measures` parameter in `GridSearchCV`, `RandomizedSearchCV`, and `cross_validate`. It's thus impossible to use those with measures that take specific parameters as input (e.g....

As discussed in #135, we can compute any similarity measure in two ways: - only considering the common (existing) ratings. This is what we're doing ATM in surprise, see the...

Replace issue templates with new issue forms

3

This PR introduces the new issue forms that replace issue templates. This is similar to what was done in torchvision https://github.com/pytorch/vision/pull/4299, you can see the end result here: https://github.com/pytorch/vision/issues/new/choose (click...

Let's say I have 2 flows and I want to compare their performances on the same benchmark suite. Is there a recommended workflow for this? My understanding is that I...

Run BaseSearchCV instances which are not scikit-learn builtins

2

#### Description I have a custom implementation of Successive Halving https://github.com/scikit-learn/scikit-learn/pull/13900 which inherits from BaseSearchCV. When I use `run_model_on_task` I'm getting the following warning: > Warning! Using subclass BaseSearchCV other...

Recommended way to shuffle intra and inter archives?

8

Say I have a bunch of archives containing samples. In my case each archive is a pickle file containing a list of samples, but it could be a tar or...

Chainer/Concater from single datapipe?

7

The `Concater` datapipe takes multiple DPs as input. Is there a class that would take a **single** datapipe of iterables instead? Something like this: ```py class ConcaterIterable(IterDataPipe): def __init__(self, source_datapipe):...

DataLoader2.__len__() ?

1

This is somewhat related to https://github.com/pytorch/data/issues/533 As described in https://github.com/pytorch/data/issues/533#issuecomment-1163381945, we like to check the `len()` of the DataLoader in torchvision in our logging utils. Are there plans to implement...