Ecosyste.ms: Issues
An open API service for providing issue and pull request metadata for open source projects.
GitHub / soda-inria/sklearn-numba-dpex issues and pull requests
#123 - WIP: Feature/upgrade dpex syntax
Pull Request -
State: open - Opened by ZzEeKkAa 9 months ago
#122 - MNT Bump dependencies to latest versions
Pull Request -
State: closed - Opened by fcharras 11 months ago
#121 - Add workarounds for https://github.com/numba/numba/issues/9242
Pull Request -
State: open - Opened by Hardcode84 about 1 year ago
#120 - MNT remove hacks after some stability issues have been fixed upstream
Pull Request -
State: closed - Opened by fcharras about 1 year ago
- 1 comment
#119 - MNT use latest numba_dpex dev version that fixes stability issues
Pull Request -
State: closed - Opened by fcharras about 1 year ago
#118 - MNT Update dependencies
Pull Request -
State: closed - Opened by fcharras about 1 year ago
- 3 comments
#117 - MNT remove done todos
Pull Request -
State: closed - Opened by fcharras about 1 year ago
#116 - MNT Minor maintenance and fix
Pull Request -
State: closed - Opened by fcharras about 1 year ago
#115 - FIX Workaround wrong compiled code on cpu
Pull Request -
State: closed - Opened by fcharras over 1 year ago
- 1 comment
#114 - MNT Synchronize with sklearn 1.4
Pull Request -
State: closed - Opened by fcharras over 1 year ago
#113 - Fix/single out 906 unit test
Pull Request -
State: closed - Opened by fcharras over 1 year ago
#112 - Fix/fix docker build pipeline
Pull Request -
State: closed - Opened by fcharras over 1 year ago
#111 - FEA kneighbors driver
Pull Request -
State: open - Opened by fcharras over 1 year ago
#110 - ENH Reuse TopK buffers
Pull Request -
State: open - Opened by fcharras over 1 year ago
- 1 comment
#109 - Fix docker build pipeline
Pull Request -
State: closed - Opened by fcharras over 1 year ago
- 8 comments
#108 - Assume private arrays are created already initialized to zero
Pull Request -
State: closed - Opened by fcharras over 1 year ago
#107 - Improve performance accross several kernels with better parameters
Pull Request -
State: closed - Opened by fcharras over 1 year ago
- 3 comments
#106 - FEA extending top-k to multi row inputs
Pull Request -
State: closed - Opened by fcharras over 1 year ago
- 1 comment
#105 - Fix monkey patching hack
Pull Request -
State: closed - Opened by fcharras over 1 year ago
- 4 comments
#104 - FIX fix performance regression seen in `numba_dpex==0.20.0dev3`
Pull Request -
State: closed - Opened by fcharras over 1 year ago
#103 - Inline comments that hint at possible future optimizations in some kernels
Pull Request -
State: open - Opened by fcharras over 1 year ago
#102 - FEA Implement an optimized sliding window matmul kernel
Pull Request -
State: closed - Opened by fcharras over 1 year ago
- 14 comments
#101 - Move reduction kernels to reduction module
Pull Request -
State: closed - Opened by fcharras over 1 year ago
- 1 comment
#100 - Hotfix strict convergence check
Pull Request -
State: closed - Opened by fcharras over 1 year ago
- 2 comments
#99 - Hotfix strict convergence check initialization
Pull Request -
State: closed - Opened by fcharras over 1 year ago
#98 - ENH Use a 2D grid of work items where applicable
Pull Request -
State: closed - Opened by fcharras over 1 year ago
- 1 comment
#97 - `test_predict_kmeans` sklearn test can sometimes fail because of non-deterministic cluster relocation
Issue -
State: open - Opened by fcharras almost 2 years ago
- 1 comment
#96 - Fix work around #906
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
- 2 comments
#95 - Measure GPU occupancy when running the benchmark on a Flex 170 on Intel DevCloud
Issue -
State: open - Opened by ogrisel almost 2 years ago
- 4 comments
#94 - Leverage the lightweight dpcpp-llvm-spirv dependency
Issue -
State: open - Opened by ogrisel almost 2 years ago
#93 - MNT bump dependencies and update, including removing hacks required by 0.19.0
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
- 1 comment
#92 - Fix conda-install mistake in readme
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
#91 - MNT Tests now rely on fallback rather than on raising NotSupportedByEngineError
Pull Request -
State: open - Opened by fcharras almost 2 years ago
#90 - ENH Add CLI to the benchmark script.
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
- 2 comments
#89 - FEAT: implement convert to sklearn types
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
- 2 comments
#88 - ENH small cleaning and optimizations accross the repo
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
- 8 comments
#87 - Improve understandability of the sum reduction kernels
Pull Request -
State: closed - Opened by ogrisel almost 2 years ago
- 3 comments
#86 - FIX make test_error_raised_on_invalid_group_sizes pass on smaller GPUs
Pull Request -
State: closed - Opened by ogrisel almost 2 years ago
- 1 comment
#85 - Fix issues seen on devcloud with flex170 gpu
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
- 2 comments
#84 - Fix a bug that caused some kmeans helper kernels to not be correctly cached
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
- 1 comment
#83 - Update dockerfile
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
- 1 comment
#82 - MAINT update dependencies to 2023 releases + numba_dpex 0.19.0
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
- 6 comments
#81 - FEA 1D Radix `topk` and application to cluster relocation in `KMeans`
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
- 9 comments
#80 - FEAT implement sum over axis0
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
- 11 comments
#79 - ENH Enhances and refactors sum kernel
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
- 2 comments
#78 - Fix sklearnex benchmark compatibility issues with 2023.0.0 release
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
#77 - Update dockerfile to up to date versions
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
#76 - Temporarily pin dpcpp runtime in conda install
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
#75 - Add kmeans_dpcpp to benchmark
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
- 5 comments
#74 - ENH Adapt to latest commits of the `feature/engines` branch
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
- 19 comments
#73 - FIX: fix dist configuration
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
#72 - MNT Overall cleaning and readability
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
#71 - Compute new_centroid_t before computing value of current inertia in verbose mode
Pull Request -
State: closed - Opened by oleksandr-pavlyk almost 2 years ago
- 1 comment
#70 - BLD pin the commit from wip-engine branch for smoother synchronization with its development
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
#69 - FIX: Prepare for dpnp 0.11.0
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
- 3 comments
#68 - ENH: Refresh the guide for installation using conda
Pull Request -
State: closed - Opened by fcharras almost 2 years ago
- 7 comments
#67 - CI Fix `pre-commit` checks
Pull Request -
State: closed - Opened by jjerphan almost 2 years ago
- 1 comment
#66 - Update README.md
Pull Request -
State: closed - Opened by preciousimo about 2 years ago
#65 - ENH: Store fitted attributes as dpt and return dpt tensors
Pull Request -
State: closed - Opened by fcharras about 2 years ago
- 7 comments
#64 - Code formatting pipeline doesn't seem to work
Issue -
State: closed - Opened by fcharras about 2 years ago
- 2 comments
#63 - fix kmeansplusplus daal benchmark
Pull Request -
State: closed - Opened by fcharras about 2 years ago
- 3 comments
#62 - FEA: engine accepts dpnp.ndarray and dpt.usm_ndarray objects as input data.
Pull Request -
State: closed - Opened by fcharras about 2 years ago
- 3 comments
#61 - Issue in evaluation of sklearn-numba-dpex with kmeans++
Issue -
State: closed - Opened by fcharras about 2 years ago
Labels: bug
#60 - Benchmark: do not report timings under 100ms precision level
Pull Request -
State: closed - Opened by ogrisel about 2 years ago
#59 - FIX Restore changes in #48 that were reverted by mistake in #55
Pull Request -
State: closed - Opened by fcharras about 2 years ago
#58 - ENH Up to date dockerfile
Pull Request -
State: closed - Opened by fcharras about 2 years ago
- 4 comments
#57 - MAINT Adapt metadata post ownership transfer
Pull Request -
State: closed - Opened by jjerphan about 2 years ago
- 1 comment
#56 - Unable to build sklearn-numba-dpex package
Issue -
State: closed - Opened by Vika-F about 2 years ago
- 12 comments
#55 - Merge https://github.com/soda-inria/sklearn-numba-dpex/pull/46 on master rather than in rng_kernels
Pull Request -
State: closed - Opened by fcharras about 2 years ago
#54 - Possible typo in `kernels.py:sum_reduction` function
Issue -
State: closed - Opened by oleksandr-pavlyk about 2 years ago
- 3 comments
#53 - Reference the numba_dpex f-layout bug
Pull Request -
State: closed - Opened by ogrisel about 2 years ago
#52 - Remove dependency on pyopencl when dpctl==0.14.0 is out
Issue -
State: closed - Opened by oleksandr-pavlyk about 2 years ago
- 5 comments
#51 - ENH Fix kernels that wasted overhead compute on out-of-bounds memory
Pull Request -
State: closed - Opened by fcharras about 2 years ago
- 4 comments
#50 - MAINT Move all input validation steps to engine
Pull Request -
State: closed - Opened by fcharras about 2 years ago
- 5 comments
#49 - ENH Performance fixes and improvements
Pull Request -
State: closed - Opened by fcharras about 2 years ago
- 8 comments
#48 - Refacto: get data conversion steps more visible
Pull Request -
State: closed - Opened by fcharras about 2 years ago
#47 - n_windows_per -> n_windows_for
Pull Request -
State: closed - Opened by fcharras about 2 years ago
#46 - Extend sum reduction kernel, add argmin reduction kernel
Pull Request -
State: closed - Opened by fcharras about 2 years ago
- 6 comments
#45 - Rng kernels
Pull Request -
State: closed - Opened by fcharras about 2 years ago
- 8 comments
#44 - Nicer configuration for test skipping if unsupported device
Pull Request -
State: closed - Opened by fcharras about 2 years ago
#43 - Module tidying
Pull Request -
State: closed - Opened by fcharras about 2 years ago
#42 - Add a module for rng kernels and kernel funcs
Pull Request -
State: closed - Opened by fcharras about 2 years ago
#41 - Small cleaning accross modules + minor lloyd fix
Pull Request -
State: closed - Opened by fcharras about 2 years ago
#40 - Use engine-style implementation for daal4py benchmark
Pull Request -
State: closed - Opened by fcharras about 2 years ago
#39 - Module tidying
Pull Request -
State: closed - Opened by fcharras about 2 years ago
#38 - Replace dpctl.tensor with dpt
Pull Request -
State: closed - Opened by fcharras about 2 years ago
#37 - FEA Implement kmeans++ with numba_dpex kernels
Pull Request -
State: closed - Opened by fcharras about 2 years ago
- 10 comments
#36 - Possible typo in "compute_euclidean_distances.py"
Issue -
State: closed - Opened by oleksandr-pavlyk about 2 years ago
- 3 comments
#35 - Possible typo in the code
Issue -
State: closed - Opened by oleksandr-pavlyk about 2 years ago
- 1 comment
#34 - bump 10.2 plus bugfix regarding dpnp_debug
Pull Request -
State: closed - Opened by fcharras about 2 years ago
#33 - Always use numerically stable expressions when computing inertia
Pull Request -
State: closed - Opened by fcharras about 2 years ago
- 5 comments
#32 - Downgrade dpnp from 0.10.2 to 0.10.1 as it causes segfaults
Pull Request -
State: closed - Opened by fcharras about 2 years ago
- 1 comment
#31 - Installing sklearn-numba-dpex
Issue -
State: open - Opened by fcharras about 2 years ago
- 1 comment
#30 - Add a workflow to automatically build/push the dev image
Pull Request -
State: closed - Opened by fcharras about 2 years ago
#29 - Auto-build the docker image using kaniko
Issue -
State: closed - Opened by fcharras about 2 years ago
- 10 comments
#28 - Rounding errors when computing inertia in KMeans
Issue -
State: closed - Opened by fcharras about 2 years ago
- 7 comments
#27 - fixup _IgnoreSampleWeight issue
Pull Request -
State: closed - Opened by fcharras about 2 years ago
#26 - Relocate centroids of empty clusters
Pull Request -
State: closed - Opened by fcharras about 2 years ago
- 4 comments
#25 - Add support for sample_weight to the KMeans driver
Pull Request -
State: closed - Opened by fcharras about 2 years ago
- 7 comments
#24 - The KMeans engine does not properly relocate clusters when it's empty
Issue -
State: closed - Opened by fcharras about 2 years ago