@@ -220,8 +220,14 @@ More details about the algorithm used is given in the following note.
220
220
221
221
222
222
Recently [23 ]_ introduced the sinkhorn divergence that build from entropic
223
- regularization to compute fast and differentiable geometric diveregnce between
224
- empirical distributions.
223
+ regularization to compute fast and differentiable geometric divergence between
224
+ empirical distributions. Note that we provide a function that compute directly
225
+ (with no need to pre compute the :code: `M ` matrix)
226
+ the sinkhorn divergence for empirical distributions in
227
+ :any: `ot.bregman.empirical_sinkhorn_divergence `. Similarly one can compute the
228
+ OT matrix and loss for empirical distributions with respectively
229
+ :any: `ot.bregman.empirical_sinkhorn ` and :any: `ot.bregman.empirical_sinkhorn2 `.
230
+
225
231
226
232
227
233
@@ -389,19 +395,21 @@ parallel.
389
395
390
396
In addition to teh speedup brought by regularization, one can also greatly
391
397
accelerate the estimation of Wasserstein barycenter when the support has a
392
- separable structure [21 ]_. In teh case of 2D images for instance one can replace
398
+ separable structure [21 ]_. In the case of 2D images for instance one can replace
393
399
the matrix vector production in teh bregman projections by convolution
394
400
operators. We provide an implementation of this algorithm in function
395
401
:any: `ot.bregman.convolutional_barycenter2d `.
396
402
397
403
.. hint ::
398
- Example of Wasserstein (:any: `ot.lp.barycenter `) and regularized wassrestein
404
+ Example of Wasserstein (:any: `ot.lp.barycenter `) and regularized Wasserstein
399
405
barycenter (:any: `ot.bregman.barycenter `) computation are available in the following examples:
400
406
401
407
- :any: `auto_examples/plot_barycenter_1D `
402
408
- :any: `auto_examples/plot_barycenter_lp_vs_entropic `
403
409
404
- Example of convolutional barycenter (:any: `ot.bregman.convolutional_barycenter2d `) computation for 2D images is available
410
+ Example of convolutional barycenter
411
+ (:any: `ot.bregman.convolutional_barycenter2d `) computation
412
+ for 2D images is available
405
413
in the following example:
406
414
407
415
- :any: `auto_examples/plot_convolutional_barycenter `
0 commit comments