Fix CPU torch.multinomial with noncontiguous prob tensor #5093
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
soumith
suggested changes
Feb 6, 2018
Contributor
soumith
left a comment
soumith
left a comment
There was a problem hiding this comment.
minor typo changes needed.
code looks good and correct
aten/src/TH/generic/THTensor.h
Outdated
| TH_API void THTensor_(expand)(THTensor *r, THTensor *tensor, THLongStorage *size); | ||
| TH_API void THTensor_(expandNd)(THTensor **rets, THTensor **ops, int count); | ||
|
|
||
| // resize* methods simplit resize the storage. So they may not retain the current data at current indices. |
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
aten/src/TH/generic/THTensor.h
Outdated
|
|
||
| // resize* methods simplit resize the storage. So they may not retain the current data at current indices. | ||
| // This is especially likely to happen when the tensor is not contiguous. In general, if you still need the | ||
| // values, unless you are doing some size and stride tricks, do not use reize*. |
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
aten/src/THC/generic/THCTensor.h
Outdated
| THC_API void THCTensor_(expand)(THCState *state, THCTensor *r, THCTensor *tensor, THLongStorage *sizes); | ||
| THC_API void THCTensor_(expandNd)(THCState *state, THCTensor **rets, THCTensor **ops, int count); | ||
|
|
||
| // resize* methods simplit resize the storage. So they may not retain the current data at current indices. |
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.
soumith
approved these changes
Feb 6, 2018
neerajprad
approved these changes
Feb 7, 2018
Contributor
neerajprad
left a comment
neerajprad
left a comment
There was a problem hiding this comment.
Thanks for the doc fixes, and adding on the unbatched params tests. Looks great!
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 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.
The CPU code previously uses
resize2dwhich does not safely preserve data when input is not contiguous. This PR changes it tounsqueeze1din both CPU and GPU functions, and add strided support for the two functions. As a result,distributions.Categoricalnow doesn't need to callcontiguous()when the probabilities are not batched.Also in this PR are
torch.multinomial,distrbutions.OneHotCategoricalanddistrbutions.Categorical, anddistrbutions.OneHotCategoricalanddistrbutions.Categoricalin unbatched mode.Fixes #5062
cc @neerajprad @alicanb for
distributions.*changes review.