Faq python different results: Difference between revisions
Jump to navigation
Jump to search
(Created page with "===Issue:=== Why am I seeing different results from ANNDL, ANNDLDA, UMAP, TSNE when using the same hyperparameters? ===Possible Solutions:=== Python's random seeding depend...") |
|||
| Line 5: | Line 5: | ||
===Possible Solutions:=== | ===Possible Solutions:=== | ||
Python's random seeding depends on the OS used to generate the necessary random components that each method uses to compute results. We fix a random state for these methods, so results should be reproducible on your machine and other machines that use the same OS. This is not guaranteed across different | Python's random seeding depends on the operating system (OS) used to generate the necessary random components that each method uses to compute results. We fix a random state for these methods, so results should be reproducible on your machine and other machines that use the same OS. This is not guaranteed across different OSes. See this issue being brought up with the creator of the '''umap-learn''' package for more details: https://github.com/lmcinnes/umap/issues/153. | ||
Latest revision as of 08:44, 26 January 2022
Issue:
Why am I seeing different results from ANNDL, ANNDLDA, UMAP, TSNE when using the same hyperparameters?
Possible Solutions:
Python's random seeding depends on the operating system (OS) used to generate the necessary random components that each method uses to compute results. We fix a random state for these methods, so results should be reproducible on your machine and other machines that use the same OS. This is not guaranteed across different OSes. See this issue being brought up with the creator of the umap-learn package for more details: https://github.com/lmcinnes/umap/issues/153.
Still having problems? Please contact our helpdesk at helpdesk@eigenvector.com