Skip to content

Conversation

@ricardoV94
Copy link
Member

@ricardoV94 ricardoV94 commented Jan 5, 2024

Description

The Softmax and LogSoftmax unfortunately don't support partial axes, so we have to fallback to Blockwise when more than one axis was used in the core case.

Related Issue

Checklist

Type of change

  • New feature / enhancement
  • Bug fix
  • Documentation
  • Maintenance
  • Other (please specify):

All types of axis arguments are supported by max_and_argmax
Also refactors shared logic for other batch axed Ops
Copy link
Member

@jessegrabowski jessegrabowski left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks good to go, who knew vectorzing softmax would require refactoring batch axis handling. The refactor looks much cleaner and more consistent across functions.

@ricardoV94 ricardoV94 merged commit 65967fe into pymc-devs:main Jan 7, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

enhancement New feature or request vectorization

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants