[pytorch] add half cauchy, half normal distributions #8411
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
fehiepsi
requested review from
apaszke,
colesbury,
ezyang,
gchanan,
soumith and
zdevito
as code owners
fritzo
approved these changes
Jun 13, 2018
torch/distributions/half_cauchy.py
Outdated
|
|
||
| def log_prob(self, value): | ||
| if self._validate_args: | ||
| self._validate_sample(value) |
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
apaszke
approved these changes
Jun 13, 2018
Contributor
apaszke
left a comment
apaszke
left a comment
There was a problem hiding this comment.
LGTM, apart from the double sample validation (in both distributions)
jpchen
approved these changes
Jun 13, 2018
Contributor
jpchen
left a comment
jpchen
left a comment
There was a problem hiding this comment.
in my experience, these half-distributions are used in hierarchical models and almost always to restrict to a positive support so i suppose it's fine
Contributor
Author
|
@apaszke The error is due to caffe2 build, which might be irrelevant to this pull request. |
apaszke
approved these changes
Jun 14, 2018
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This pull request adds HalfNormal, HalfCauchy for distributions. These distributions are transformed distributions with AbsTransform but AbsTransform does not have
signandlog_abs_det_jacobianmethods. So I think that it is good to implement them.As mentioned in this user request, these half distributions are popular for prior of a positive variable (currently, I guess Exponential is the only way to set prior to such variable). Missing these distributions add some overhead for benchmarking against other frameworks.
cc @apaszke @fritzo