segmentation fault issue depending on hyperparameter

Nobody has claimed this yet.

Hello.

I was facing weird occasions where UMAP works perfectly for a set of hyperparameter while another set of hyperparameter causing error exhibiting 'segmentation fault'. It is not due to data itself nor virtualenv since they were identical. FYI, data is about (500,000 x 39) numpy array.

The versions of packages are:
- umap-learn == 0.4.0 (but was also same in 0.3.0 and 0.4.0rc3).
- numba ==0.48.0 (I tried 0.40.0, 0.44.0, 0.46.0 but the result was identical)
- scikit-learn ==0.22.0
- numpy==1.18.2
- scipy == 1.4.0
- llvmlite == 0.31.0
- joblib==0.14.1
- python=3.6.1 (I tested with python=3.6.4 also)

The functioning hyperparameter set is:
- n_neighbor: 30, 50 (I did not search all yet)
- min_dist: 0.005 to 0.5
- metric: correlation
- dimension:2

The malfunctioning hyperparameter set it
-n_neighbor: 70, 90, 110 (I did not search all yet)
-min_dist: 0.005 to 0.5 (these are identical from above)
-metric: correlation
-dimension:2

Just to add, the OS system is Linux Mint 18.4. I tested both on the command line by .py and Jupyter lab.

If there are any guess why it might happen, please let me know. Thank you.

Start by reproducing the issue from the command-line .py example using the reported 500,000 × 39 NumPy array and compare the working and failing n_neighbor values. Check whether the same behavior occurs in Jupyter Lab, then narrow down the parameter combination and environment involved. Done means the segmentation fault is explained and a verified fix or reproducible diagnosis is documented.

Written by the indexing model from the issue text.