saymrwulf/pytorch
1
0
Fork 0
mirror of https://github.com/saymrwulf/pytorch.git synced 2026-05-15 21:00:47 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions
3344 commits 5021 branches 1206 tags 1 GiB
Commit graph

12 commits

Author SHA1 Message Date
Joseph Spisak
11a736b682 Sqrt op (#2101) 
* First attempt on sqrt op

* Adding the Sqrt op along with the test cases

* Made changes per @Yangqing's questions re: tensor format and used hypothesis to generate input tensor
 2018-03-02 16:19:45 -08:00
Orion Reblitz-Richardson
ccea6924a2 Implementing Pow operator (this merges existing pow with a scalar and new pow with a tensor exponent). Second Try. 
The old pow operator has been deleted in math_ops.cc, math_ops.cu and math_ops.h, while the new operator supporting scalar and tensor exponent has been added in pow_op.cc, pow_op.h an elementwise_op.cu.
 2018-02-21 18:31:45 -08:00
Pieter Noordhuis
52fa742c51 Revert D6893040: Implementing Pow operator (this merges existing pow with a scalar and new pow with a tensor exponent). 
Summary:
This reverts commit 30f614beea6f859fee25ce4f85573142885dde45

bypass-lint

An infra SEV is better than not reverting this diff.
If you copy this password, see you in SEV Review!
cause_a_sev_many_files

Differential Revision:
D6893040

Original commit changeset: 30f614beea6f

fbshipit-source-id: 5e98a24699088283f864efe31234874bdacbe3c3
 2018-02-14 10:34:08 -08:00
Maxim Naumov
f7cc8e8822 Implementing Pow operator (this merges existing pow with a scalar and new pow with a tensor exponent). 
Summary: The old pow operator has been deleted in math_ops.cc, math_ops.cu and math_ops.h, while the new operator supporting scalar and tensor exponent has been added in pow_op.cc, pow_op.h an elementwise_op.cu.

Reviewed By: houseroad

Differential Revision: D6893040

fbshipit-source-id: 30f614beea6f859fee25ce4f85573142885dde45
 2018-02-13 17:46:35 -08:00
Manoj Krishnan
6d32e36682 Caffe2 Operator: GPU implementation of Swish Activation 
Summary: GPU (CUDA) implementation of the Swish activation function in Caffe2.

Reviewed By: Yangqing, xianjiec

Differential Revision: D6656907

fbshipit-source-id: f5f2c667055abf679728d2b5d43998895ddec708
 2018-01-05 12:04:25 -08:00
Pieter Noordhuis
348e29c49b Don't run CUDA tests for ops without CUDA implementation 
Summary: Closes https://github.com/caffe2/caffe2/pull/1434

Reviewed By: houseroad, ilia-cher

Differential Revision: D6272614

Pulled By: pietern

fbshipit-source-id: 7b998b08ec02b03f88a6fd24a949b0d199b2aa37
 2017-11-08 10:28:02 -08:00
Badri Narayan Bhaskar
25bfffeafe Swish Activation Function 
Summary:
Swish: A self-gated activation function.
https://arxiv.org/pdf/1710.05941.pdf

Reviewed By: ajtulloch

Differential Revision: D6100424

fbshipit-source-id: 0103d6d82e9ffb50106c98a8785e62b8808e9af1
 2017-10-20 10:37:43 -07:00
Yangqing Jia
8286ce1e3a Re-license to Apache 
Summary: Closes https://github.com/caffe2/caffe2/pull/1260

Differential Revision: D5906739

Pulled By: Yangqing

fbshipit-source-id: e482ba9ba60b5337d9165f28f7ec68d4518a0902
 2017-09-28 16:22:00 -07:00
James Reed
85f1d947dd Vectorize SigmoidOp on CPU 
Summary: I noticed that Sigmoid was taking an inordinate amount of time in our NMT benchmark, so I looked at the implementation and it didn't seem optimal. I replaced the implementation with an Eigen version so that when the Eigen update goes through, we will get proper AVX(2) vectorization.

Differential Revision: D5082464

fbshipit-source-id: aa951f7d730fc05198f7dd04076ec58d471b74c8
 2017-05-17 20:33:36 -07:00
Aapo Kyrola
8fab453863 Sqr op and gradient 
Summary: Due to popular demand, added an op to compute element-wise square + gradient for it (just for the fun of it).

Reviewed By: Yangqing

Differential Revision: D4664797

fbshipit-source-id: 0a29c7c249fdc72f51412bebd6ae352a7801cf05
 2017-03-07 03:03:07 -08:00
Aapo Kyrola
8caa7cec8d CUDA version of Log 
Summary: As in the title. Simple registration issue.

Reviewed By: Yangqing, jhcross

Differential Revision: D4655691

fbshipit-source-id: 661e4d5f1226ec05e099c84f4454aa07c6be4449
 2017-03-04 00:32:03 -08:00
Aapo Kyrola
4f1db36cff add CUDA gradient for Div 
Summary: DivOp missed a gradient for CUDA, so implemented it. Also added operator test.

Differential Revision: D4396638

fbshipit-source-id: 9949e47aa3735bb418a0db003e2b2f4896056a71
 2017-01-09 21:59:23 -08:00