Saiteja Utpala
Saiteja Utpala
### Where should the content be modified? https://github.com/geomstats/geomstats/blob/master/geomstats/geometry/poincare_polydisk.py ### What should be modified? Poincare Polydisk metric should be inherited from NFoldMetric ### Additional information _No response_
This PR collects and adds properties to manifolds and metrics. addressing #1475
### Where should the content be modified? https://github.com/geomstats/geomstats/blob/master/geomstats/geometry/poincare_polydisk.py ### What should be modified? Poincare Polydisk should be inherited from NFold Manifold ### Additional information _No response_
### Where should the content be modified? geomstats/_backend ### What should be modified? using geomstats with `pytorch` on `gpu` is failing on several things because in pytorch one has to...
### Where should the content be modified? https://github.com/geomstats/geomstats/blob/master/geomstats/geometry/special_euclidean.py ### What should be modified? currently `SpecialEuclideanMatrixCannonicalLeftMetric` takes `group` as argument. But it should take parameters of group instead of group itself....
### Where should the content be modified? https://github.com/geomstats/geomstats/blob/master/geomstats/geometry/special_orthogonal.py ### What should be modified? create separate metric class inside special orthogonal which takes `dim`, `point_type`. Currently it is being set directly....
bp = SPDMatrices(1).random_point(1) SPDMetricAffine(1).log(bp, bp)
spd_matrices projection uses gs.atol. So output is varied based on different backend (although by small amount). I think it's better if we have same output across backends for simple operations...
``` import numpy as np from geomstats.geometry.matrices import Matrices Matrices.equal(np.eye(3), np.eye(2)) ``` should return false