Julien Jerphanion
Julien Jerphanion
#### Reference Issues/PRs Part of https://github.com/scikit-learn/scikit-learn/issues/22587. To be merged after #24120. #### What does this implement/fix? Explain your changes. Currently, the heuristic to choose between `parallel_on_X` or `parallel_on_Y` is sub-optimal...
#### Reference Issues/PRs Relates to #22587 #### What does this implement/fix? Explain your changes. I think investing some time in coming up with a proper benchmarking suite for `PairwiseDistancesReductions` is...
#### Reference Issues/PRs Follows-up https://github.com/scikit-learn/scikit-learn/pull/22134 #### What does this implement/fix? Explain your changes. This ports `PairwiseDistancesReduction` and other implementations to 32bit using Tempita. #### Benchmarks results The hardware scalability plateaus...
#### Reference Issues/PRs Comes after https://github.com/scikit-learn/scikit-learn/pull/23515. Relates to https://github.com/scikit-learn/scikit-learn/issues/22587. #### What does this implement/fix? Explain your changes. Add `SparseSparseDatasetsPair`, `SparseDenseDatasetsPair`, `DenseSparseDatasetsPair` to bridge distances computations for pairs of fused dense-sparse...
#### What does this implement/fix? Explain your changes. NumPy has several C API versions. Currently, scikit-learn is using the oldest one. This currently might not offer the best performance and...
`PairwiseDistancesReductions` have been introduced as a hierarchy of Cython classes to implement back-ends of some scikit-learn algorithms. Pieces of work include: - https://github.com/scikit-learn/scikit-learn/pull/22134 - https://github.com/scikit-learn/scikit-learn/pull/22320 - https://github.com/scikit-learn/scikit-learn/pull/23865 - Support sparse-dense...
Hello, @h-sdl and I have implemented a small pipeline for change detection for polarimetric SAR images in python from [1]. In short, we can detect changes between two SAR and...
### Context When creating a [`scipy.sparse.csr_matrix`](https://docs.scipy.org/doc/scipy/reference/generated/scipy.sparse.csr_matrix.html) (for instance using a numpy array), the underlying `indices` and `indptr` arrays either use `np.int32` or `np.int64` depending on `nnz`, the number of non...
Hello, CGAL does have a few fixtures internally to [compute mesh statistics](https://doc.cgal.org/latest/Polygon_mesh_processing/index.html#fig__Fig_smooth_stats). I think having an interface in PyMesh would be a nice feature. I started to work on a...
Being able to specify [`skinparams`](https://plantuml.com/skinparam) as well as class diagram syntactic sugar for [layout and arrow direction options](https://plantuml.com/class-diagram) for the output might be helpful. Integrating it in the python API...