Ecosyste.ms: Timeline

Browse the timeline of events for every public repo on GitHub. Data updated hourly from GH Archive.

Pierre-Bartet

Pierre-Bartet forked dask/dask

Pierre-Bartet/dask

Pierre-Bartet created a comment on an issue on dask/dask
Yes there are some aggregation then dispatching to be done but it is the case for many algorithms.

View on GitHub

Pierre-Bartet created a comment on an issue on dask/dask
Nice, so then what is the limitation preventing a true lazy implementation?

View on GitHub

Pierre-Bartet created a comment on an issue on dask/dask
Ok thanks for explanation. Indeed now I understand why I was losing my sanity ^^ Is this non lazy sorting also non parallel (does `dask` squashes the chunks before sorting)?

View on GitHub

Pierre-Bartet created a comment on an issue on dask/dask
I still don't get it, using `df.set_index("some_colulmn")` doesn't immediately trigger a computation, so it still look like at least somewhat lazy, yet I can't find a parallel merge sort happening ...

View on GitHub

Pierre-Bartet created a comment on an issue on dask/dask
Thanks I was questioning my own sanity!

View on GitHub

Pierre-Bartet created a comment on an issue on dask/dask
Indeed, so could we do the same thing for `dask` arrays as what is done for `dask` dataframes `set_index`? Or am I missing something (for example `set_index` also squashing the chunk before sorti...

View on GitHub

Pierre-Bartet created a comment on an issue on dask/dask
I gave a look and it looks like it's still hard not in itself but because if I understand properly the existing shuffle (such as `set_index`) are still not lazy (which is what stopped me 6 years ag...

View on GitHub

Pierre-Bartet created a comment on an issue on dask/dask
Interesting to see that this is still open 6 years later! I still find the solution to silently squash chunks very treacherous. Countless people will be bitten by such an unexpected behavior. As ...

View on GitHub

Pierre-Bartet created a comment on a pull request on onnx/sklearn-onnx
@xadupre done

View on GitHub

Pierre-Bartet created a comment on a pull request on onnx/onnx
Perfect then

View on GitHub

Pierre-Bartet created a comment on a pull request on onnx/onnx
> it still needs to "always contain UTF-8 encoded or 7-bit ASCII text, and cannot be longer than 2^32." according to the protobuf string type definition. Sounds reasonable, and then the checker ...

View on GitHub

Pierre-Bartet pushed 1 commit to add_select_k_best_test Pierre-Bartet/sklearn-onnx
  • Rely on a public sklearn attribute instead of a private one a4af89e

View on GitHub

Pierre-Bartet created a comment on a pull request on onnx/sklearn-onnx
No `tests/test_sklearn_feature_selection_converters.py` test fail when doing that, but other tests fail elsewhere. So from what I understand, the answer to question 2. is "Yes". I guess this is a...

View on GitHub

Pierre-Bartet pushed 1 commit to add_select_k_best_test Pierre-Bartet/sklearn-onnx

View on GitHub

Pierre-Bartet created a comment on a pull request on onnx/sklearn-onnx
@xadupre I fixed the issue (just for `SelectKBest`, not the other selectors), but before I go further I want to be sure I understand properly what was the initial intention of the code I changed. ...

View on GitHub

Pierre-Bartet pushed 7 commits to add_select_k_best_test Pierre-Bartet/sklearn-onnx

View on GitHub

Pierre-Bartet pushed 1 commit to add_select_k_best_test Pierre-Bartet/sklearn-onnx

View on GitHub

Pierre-Bartet created a comment on a pull request on onnx/sklearn-onnx
This is a one liner fix and all the existing and added tests pass so yes!

View on GitHub

Pierre-Bartet created a comment on a pull request on onnx/sklearn-onnx
Sorry for not running Black before pushing then ^^

View on GitHub

Pierre-Bartet created a comment on a pull request on onnx/sklearn-onnx
Are the workflows run automatically when pushing a new commit or do you have to approve each run each time ?

View on GitHub

Pierre-Bartet pushed 3 commits to fix_empty_column_selector Pierre-Bartet/sklearn-onnx

View on GitHub

Pierre-Bartet pushed 1 commit to fix_empty_column_selector Pierre-Bartet/sklearn-onnx

View on GitHub

Pierre-Bartet created a comment on a pull request on onnx/sklearn-onnx
@xadupre this one is ready for review (and I turned the non ready one into a draft).

View on GitHub

Pierre-Bartet created a comment on an issue on onnx/sklearn-onnx
Yes they never actually enforced it, and now it has even been removed from the specification (from MUST to SHOULD). Does it mean `sklearn-onnx` can just drop the corresponding check now?

View on GitHub

Pierre-Bartet created a comment on an issue on onnx/sklearn-onnx
The `onnx` specification has just been updated to drop the C90 identifier requirement: https://github.com/onnx/onnx/pull/6652.

View on GitHub

Pierre-Bartet created a comment on a pull request on onnx/onnx
Nice, does it mean that the names can be any string of bytes, or is it still restricted to some subset such as ascii ?

View on GitHub

Pierre-Bartet pushed 6 commits to fix_empty_column_selector Pierre-Bartet/sklearn-onnx

View on GitHub

Pierre-Bartet pushed 2 commits to fix_empty_column_selector Pierre-Bartet/sklearn-onnx

View on GitHub

Pierre-Bartet opened a pull request on onnx/sklearn-onnx
Fix empty column selector
Fixes #1155 by skipping transforms with no corresponding columns at all, and add the corresponding test.
Load more