Incubator-mxnet: Numpy Diag in MXNet

Created on 29 Dec 2017  路  4Comments  路  Source: apache/incubator-mxnet

Call for Contribution Operator

Most helpful comment

I am working on replicating np.diag(). Once done will do the full diagonal operator. This is my first operator so I am learning a lot at the moment.

  • [x] 2d forward pass CPU
  • [x] 2d forward pass GPU
  • [x] 1d forward pass CPU
  • [x] 1d forward pass GPU
  • [x] 2d backward pass CPU
  • [x] 2d backward pass GPU
  • [x] 1d backward pass CPU
  • [x] 1d backward pass GPU
  • [x] unit tests

PR:
https://github.com/apache/incubator-mxnet/pull/11643

All 4 comments

I am interested implementing this, is there a jira issue already?

@ifeherva NoJIRA issue for this one yet.

I am working on replicating np.diag(). Once done will do the full diagonal operator. This is my first operator so I am learning a lot at the moment.

  • [x] 2d forward pass CPU
  • [x] 2d forward pass GPU
  • [x] 1d forward pass CPU
  • [x] 1d forward pass GPU
  • [x] 2d backward pass CPU
  • [x] 2d backward pass GPU
  • [x] 1d backward pass CPU
  • [x] 1d backward pass GPU
  • [x] unit tests

PR:
https://github.com/apache/incubator-mxnet/pull/11643

@eric-haibin-lin This issue can be closed now that the PR has been merged

Was this page helpful?
0 / 5 - 0 ratings

Related issues

sbodenstein picture sbodenstein  路  3Comments

phunterlau picture phunterlau  路  3Comments

Ajoo picture Ajoo  路  3Comments

seongkyun picture seongkyun  路  3Comments

realbns2008 picture realbns2008  路  3Comments