data-science
Here are 14,329 public repositories matching this topic...
Describe the issue linked to the documentation
Follows from #17387
Suggest a potential alternative/fix
Stop referencing preprocessing functions e.g. :
maxabs_scale
minmax_scale
normalize
quantile_transform
robust_scale
scale
power_transform
in the UG, and only add them e.g. in the "See Also" sections, or even just in the API ref.
In particular right now the first entr
This issue tracks Superset's migration from JavaScript to TypeScript (as started in SIP-36). If you'd like to help with the migration, feel free to take an unchecked directory and convert the files within the immediate directory from JavaScript/JSX to TypeScript/TSX. #9162 and #9180 provide some tips for performing the migration. Once com
-
Updated
Aug 12, 2020 - Jupyter Notebook
-
Updated
Jul 24, 2020 - Python
-
Updated
Sep 4, 2020 - Python
-
Updated
Aug 17, 2020 - Python
There are many places where we could make use of Pathlib.
Look for any places that uses with open(...) and ask yourself:
- is the argument a string ?
- would it make sens to make it a
Path(), - how far upstream in the code can I make it a Path.
Don't try to bite more than you can chew (or more than I can review), try to fix 1 place at a time.
-
Updated
Aug 11, 2020
Describe your feature request
Currently, if you run
ray.wait([pg.ready()]), it prints the warning message that the ready task is infeasible. This is an expected behavior, so we should not print warning messages.
Impl Idea
- Have a magic number for pg.ready resources (for example, it always re
-
Updated
Sep 3, 2020 - Python
The usage example in the word2vec.py doc-comment regarding KeyedVectors uses inconsistent paths and thus doesn't work.
If vectors were saved to a tm
-
Updated
Aug 10, 2020 - Jupyter Notebook
-
Updated
May 20, 2020
Summary
The following maps to https://twemoji.maxcdn.com/2/72x72/1f6f0-fe0f.png, which 404s.
st.beta_set_page_config(
page_icon="🛰️",
)
# yields
# <link rel="shortcut icon" href="https://twemoji.maxcdn.com/2/72x72/1f6f0-fe0f.png">As a result, the custom favicon does not load.
Steps to reproduce
See above.
Expected behavior:
I expect the `:sa
-
Updated
Aug 20, 2020
-
Updated
Jul 31, 2020
-
Updated
May 10, 2020 - Python
-
Updated
Aug 24, 2020
more details at: allenai/allennlp#2264 (comment)
-
Updated
Sep 3, 2020 - Jupyter Notebook
-
Updated
Sep 1, 2020 - Python
-
Updated
Aug 18, 2020 - Jupyter Notebook
Hand-chosen values are not enough, we need to test with a large batch of inputs where possible.
Something in this style, maybe with a fixed seed:
def test_auroc_versus_sklearn():
for i in range(100):
target = torch.randint(0, 2, size=(10, ))
pred = torch.randint(0, 2, size=(10,))
score_sk = sklearn.metrics.roc_auc_score(target.numpy(), pred.numpy()) -
Updated
Sep 3, 2020 - Python
Is your feature request related to a problem or area of OpenRefine? Please describe.
When previewing an importer's results, there is a text field to enter the project name. When pressing Enter in this field, I would expect the project to be created (without having to click the "Create Project" button).
, because some of these examples are now being rendered in the docs.
Added by @fchollet in requests for contributions.