[REVIEW] Adding initial artifacts for umap paper reproducible benchmarks

publications/umap_paper_benchmarks/README.md

@@ -0,0 +1,34 @@
+Datasets are not included in this repository and need to be downloaded separately. 
+
+The necessary dependencies for reproducing the benchmarks have been captured in `conda` environment yaml files. 
+
+To install dependencies for cuml and UMAP-learn benchmarks: 
+```
+conda env create --name cuml_umap_benchmarks -f conda/umap_paper_cuml_cuda10.2.yml
+```
+
+To install dependencies for GPUMAP benchmarks: 
+```
+conda env create --name gpumap_benchmarks -f conda/umap_paper_gpumap_cuda10.0.yml
+```
+
+You can run the notebooks using jupyter lab:
+```
+conda activate <environment_name>
+python -m ipykernel install --user
+jupyter lab
+```
+
+# Datasets
+
+- PEN Digits - uses sklearn.datasets.load_digits
+- GoogleNews Word2Vec - Downloaded from https://code.google.com/archive/p/word2vec/ and loaded using Gensim library
+- Fashion MNIST - Downloaded from https://github.com/zalandoresearch/fashion-mnist
+- CIFAR-100 - Downloaded from https://www.cs.toronto.edu/~kriz/cifar.html
+- Shuttle - Downloaded from https://archive.ics.uci.edu/ml/datasets/Statlog+(Shuttle)
+- MNIST - Uses datasets submodule to download and load
+- TASIC2018 - Data from : https://portal.brain-map.org/atlases-and-data/rnaseq (see dedicated notebook)
+- scRNA - Dataset downloaded from https://cells.ucsc.edu/
+- COIL-20 - Uses datasets submodule to download and load
+
+

publications/umap_paper_benchmarks/conda/umap_paper_cuml_cuda10.2.yml

@@ -0,0 +1,37 @@
+channels:
+  - rapidsai-nightly
+  - nvidia
+  - conda-forge
+  - facebook
+  - pytorch
+  - defaults
+dependencies:
+  - python=3.7
+  - cudatoolkit=10.2
+  - cudf
+  - dask-cuda
+  - dask-cudf
+  - cuml
+  - cugraph
+  - opencv
+  - scikit-image
+  - chainer
+  - scipy
+  - gensim
+  - ucx-py
+  - joblib
+  - matplotlib
+  - umap-learn
+  - numba
+  - ucx-proc=*=gpu
+  - scikit-learn
+  - cupy
+  - ipykernel
+  - jupyterlab
+  - pip
+  - pip:
+      - jupyter-server-proxy
+      - git+https://github.com/dask/dask.git
+      - git+https://github.com/dask/distributed.git
+      - git+https://github.com/overshiki/datasets.git
+      - wget

publications/umap_paper_benchmarks/conda/umap_paper_gpumap_cuda10.0.yml

@@ -0,0 +1,39 @@
+channels:
+  - rapidsai-nightly
+  - nvidia
+  - conda-forge
+  - facebook
+  - pytorch
+  - defaults
+dependencies:
+  - python=3.7
+  - cudatoolkit=10.0
+  - cudf
+  - dask-cuda
+  - dask-cudf
+  - cuml
+  - cugraph
+  - opencv
+  - scikit-image
+  - chainer
+  - scipy
+  - ucx-py
+  - joblib
+  - matplotlib
+  - umap-learn
+  - numba
+  - ucx-proc=*=gpu
+  - scikit-learn
+  - cupy
+  - ipykernel
+  - jupyterlab
+  - pip
+  - pip:
+      - jupyter-server-proxy
+      - faiss==1.5.3
+      - faiss-gpu==1.5.3
+      - git+https://github.com/dask/dask.git
+      - git+https://github.com/dask/distributed.git
+      - git+https://github.com/p3732/gpumap.git
+      - git+https://github.com/overshiki/datasets.git
+      - wget