Skip to content

[MRG] Fix#421 pass stopThr to the sinkhorn function in empirical_sinkhorn_divergence #422

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 4 commits into from
Dec 21, 2022
Merged

[MRG] Fix#421 pass stopThr to the sinkhorn function in empirical_sinkhorn_divergence #422

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
32101c7
fix stopThr hardcoded in some places
tlacombe Dec 20, 2022
ab49b3d
added fix documentation in RELEASES.Md
tlacombe Dec 20, 2022
0f0a3e5
Merge branch 'master' into fix421
rflamary Dec 20, 2022
4c11591
Merge branch 'master' into fix421
rflamary Dec 20, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions RELEASES.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -32,6 +32,8 @@ roughly 2^31) (PR #381)
- Fixed weak optimal transport docstring (Issue #404, PR #410)
- Fixed error whith parameter `log=True`for `SinkhornLpl1Transport` (Issue #412,
PR #413)
- Fix an issue where the parameter `stopThr` in `empirical_sinkhorn_divergence` was rendered useless by subcalls
that explicitly specified `stopThr=1e-9` (Issue #421, PR #422).
- Fixed a bug breaking an example where we would try to make an array of arrays of different shapes (Issue #424, PR #425)


Expand Down
14 changes: 7 additions & 7 deletions ot/bregman.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1281,7 +1281,7 @@ def get_reg(n): # exponential decreasing
regi = get_reg(ii)

G, logi = sinkhorn_stabilized(a, b, M, regi,
numItermax=numInnerItermax, stopThr=1e-9,
numItermax=numInnerItermax, stopThr=stopThr,
warmstart=(alpha, beta), verbose=False,
print_period=20, tau=tau, log=True)

Expand Down Expand Up @@ -3306,17 +3306,17 @@ def empirical_sinkhorn_divergence(X_s, X_t, reg, a=None, b=None, metric='sqeucli
if log:
sinkhorn_loss_ab, log_ab = empirical_sinkhorn2(X_s, X_t, reg, a, b, metric=metric,
numIterMax=numIterMax,
stopThr=1e-9, verbose=verbose,
stopThr=stopThr, verbose=verbose,
log=log, warn=warn, **kwargs)

sinkhorn_loss_a, log_a = empirical_sinkhorn2(X_s, X_s, reg, a, a, metric=metric,
numIterMax=numIterMax,
stopThr=1e-9, verbose=verbose,
stopThr=stopThr, verbose=verbose,
log=log, warn=warn, **kwargs)

sinkhorn_loss_b, log_b = empirical_sinkhorn2(X_t, X_t, reg, b, b, metric=metric,
numIterMax=numIterMax,
stopThr=1e-9, verbose=verbose,
stopThr=stopThr, verbose=verbose,
log=log, warn=warn, **kwargs)

sinkhorn_div = sinkhorn_loss_ab - 0.5 * (sinkhorn_loss_a + sinkhorn_loss_b)
Expand All @@ -3333,17 +3333,17 @@ def empirical_sinkhorn_divergence(X_s, X_t, reg, a=None, b=None, metric='sqeucli

else:
sinkhorn_loss_ab = empirical_sinkhorn2(X_s, X_t, reg, a, b, metric=metric,
numIterMax=numIterMax, stopThr=1e-9,
numIterMax=numIterMax, stopThr=stopThr,
verbose=verbose, log=log,
warn=warn, **kwargs)

sinkhorn_loss_a = empirical_sinkhorn2(X_s, X_s, reg, a, a, metric=metric,
numIterMax=numIterMax, stopThr=1e-9,
numIterMax=numIterMax, stopThr=stopThr,
verbose=verbose, log=log,
warn=warn, **kwargs)

sinkhorn_loss_b = empirical_sinkhorn2(X_t, X_t, reg, b, b, metric=metric,
numIterMax=numIterMax, stopThr=1e-9,
numIterMax=numIterMax, stopThr=stopThr,
verbose=verbose, log=log,
warn=warn, **kwargs)

Expand Down