Add float16 support to dropout operator #9223
Merged
Conversation
This file contains 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
chengduoZH
reviewed
Mar 20, 2018
|
|
||
| def test_check_output(self): | ||
| if core.is_compiled_with_cuda() and core.op_support_gpu("dropout"): | ||
| self.check_output_with_place(core.CUDAPlace(0), atol=1e-3) |
There was a problem hiding this comment.
TestFP16DropoutOp1 and TestFP16DropoutOp2 are very similar, and they can be inherited relationships, which can reduce the amount of code.
There was a problem hiding this comment.
You are right! Done.
chengduoZH
approved these changes
Mar 20, 2018
| #else | ||
| HOSTDEVICE inline float16 operator+(const float16& a, const float16& b) { | ||
| HOST inline float16 operator+(const float16& a, const float16& b) { |
There was a problem hiding this comment.
Maybe HOST is unnecessary.
There was a problem hiding this comment.
I guess so. Let me fix that in the next PR.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
fix #9222
Added device function for multiplying two float16 numbers on GPU device, which is needed in the following code: