0.2.9: new initializations for LR Sinkhorn and LR GW (including k-means++ implementations), entropic map

[marcocuturi]

What's Changed

• Feature/kmeans++ by @michalk8 in Feature/kmeans++ #120
• Feature/ Introduce initialization methods for Sinkhorn by @JTT94 in Feature/ Introduce initialization methods for Sinkhorn #98
• Pin sphinx-book-theme>=0.3.3 by @michalk8 in Pin sphinx-book-theme>=0.3.3 #123
• Additional Docstrings and Comments to NeuralDual-Module by @bunnech in Additional Docstrings and Comments to NeuralDual-Module #122
• Allow k-means to be differentiable when using while loop by @michalk8 in Allow k-means to be differentiable when using while loop #130
• Feature/graph geometry by @michalk8 in Feature/graph geometry #126
• LR Sinkhorn improvements by @meyerscetbon in LR Sinkhorn improvements #111
• Refactor GW initialization by @michalk8 in Refactor GW initialization #133
• Change LR initializer to random by @michalk8 in Change LR initializer to random #134
• Generalized k-means initialization for LR Sinkhorn/Gromov-Wasserstein by @michalk8 in Generalized k-means initialization for LR Sinkhorn/Gromov-Wasserstein #135
• Switch to one linear system per implicit differentiation by @ersisimou in Switch to one linear system per implicit differentiation #136
• Feature/disable warm start by @michalk8 in Feature/disable warm start #138
• Add entropic map by @michalk8 in Add entropic map #142

Full Changelog: 0.2.8...0.2.9

---

This discussion was created from the release 0.2.9: new initializations for LR Sinkhorn and LR GW (including k-means++ implementations), entropic map.