search

ramsey-devs / ramsey

Probabilistic deep learning using JAX
https://ramsey.rtfd.io
Apache License 2.0
13 stars 3 forks source link

feat: added SparseGP model and example #21

Closed mamarder closed 2 years ago

mamarder commented 2 years ago

Hello @dirmeier,

I implemented a Sparse GP model according Titsias, Michalis paper Variational Learning of Inducing Variables in Sparse Gaussian Processes from April 2009.

Further I added an example which demonstrates how to set up and train the model as well as how to visualizing the data.

See the picture below for a first impression of the added example.

image

Please review.

Regards Marco

codecov[bot] commented 2 years ago

Codecov Report

Base: 50.75% // Head: 45.62% // Decreases project coverage by -5.12% :warning:

Coverage data is based on head (28088fa) compared to base (aa857d6). Patch coverage: 19.00% of modified lines in pull request are covered.

Additional details and impacted files ```diff @@ Coverage Diff @@ ## main #21 +/- ## ========================================== - Coverage 50.75% 45.62% -5.13% ========================================== Files 21 22 +1 Lines 532 629 +97 ========================================== + Hits 270 287 +17 - Misses 262 342 +80 ``` | [Impacted Files](https://codecov.io/gh/ramsey-devs/ramsey/pull/21?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=ramsey-devs) | Coverage Δ | | |---|---|---| | [ramsey/\_src/gaussian\_process/gaussian\_process.py](https://codecov.io/gh/ramsey-devs/ramsey/pull/21/diff?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=ramsey-devs#diff-cmFtc2V5L19zcmMvZ2F1c3NpYW5fcHJvY2Vzcy9nYXVzc2lhbl9wcm9jZXNzLnB5) | `29.26% <ø> (-0.97%)` | :arrow_down: | | [...ey/\_src/gaussian\_process/train\_gaussian\_process.py](https://codecov.io/gh/ramsey-devs/ramsey/pull/21/diff?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=ramsey-devs#diff-cmFtc2V5L19zcmMvZ2F1c3NpYW5fcHJvY2Vzcy90cmFpbl9nYXVzc2lhbl9wcm9jZXNzLnB5) | `18.00% <4.76%> (-9.59%)` | :arrow_down: | | [...y/\_src/gaussian\_process/sparse\_gaussian\_process.py](https://codecov.io/gh/ramsey-devs/ramsey/pull/21/diff?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=ramsey-devs#diff-cmFtc2V5L19zcmMvZ2F1c3NpYW5fcHJvY2Vzcy9zcGFyc2VfZ2F1c3NpYW5fcHJvY2Vzcy5weQ==) | `21.79% <21.79%> (ø)` | | | [ramsey/train.py](https://codecov.io/gh/ramsey-devs/ramsey/pull/21/diff?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=ramsey-devs#diff-cmFtc2V5L3RyYWluLnB5) | `100.00% <100.00%> (ø)` | | Help us with your feedback. Take ten seconds to tell us [how you rate us](https://about.codecov.io/nps?utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=ramsey-devs). Have a feature suggestion? [Share it here.](https://app.codecov.io/gh/feedback/?utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=ramsey-devs)

:umbrella: View full report at Codecov.
:loudspeaker: Do you have feedback about the report comment? Let us know in this issue.