Skip to content

GitLab

  • Menu
Projects Groups Snippets
    • Loading...
  • Help
    • Help
    • Support
    • Community forum
    • Submit feedback
    • Contribute to GitLab
  • Sign in / Register
  • ketos ketos
  • Project information
    • Project information
    • Activity
    • Labels
    • Planning hierarchy
    • Members
  • Repository
    • Repository
    • Files
    • Commits
    • Branches
    • Tags
    • Contributors
    • Graph
    • Compare
  • Issues 27
    • Issues 27
    • List
    • Boards
    • Service Desk
    • Milestones
  • Merge requests 0
    • Merge requests 0
  • CI/CD
    • CI/CD
    • Pipelines
    • Jobs
    • Schedules
  • Deployments
    • Deployments
    • Environments
    • Releases
  • Monitor
    • Monitor
    • Incidents
  • Analytics
    • Analytics
    • Value stream
    • CI/CD
    • Repository
  • Wiki
    • Wiki
  • Snippets
    • Snippets
  • Activity
  • Graph
  • Create a new issue
  • Jobs
  • Commits
  • Issue Boards
Collapse sidebar
  • public_projects
  • ketosketos
  • Issues
  • #132

Closed
Open
Created Nov 04, 2021 by Fabio Frazao@fsfrazaoOwner

"ketos-run" command line tool

I suggest that we add a command line tool to ketos, which could be used to run trained models.

The idea is similar to the "detector.py" script mentioned in the Training a detector tutorial, but it would be installed with ketos and therefore available for users to directly call like:

$ ketos-run my_model.kt --data=my/wav/files ...

However, it would be useful if this command-line tool were more general: you pass the .kt model and the data, and it does what the model developer intended it to do. And to that end, the .kt model could include a new settings file (maybe "inference.json"), which would basically provide the necessary instructions to the "ketos-run" tool (is this a detector supposed to run on continuous .wav files or just a classifier expecting short clips? What are the default parameters?).

I think this would contribute to making trained ketos models easier to use. If someone gives you a trained detector, you just have to install ketos, call that command on your data and it will do what the model developer intended it to do. You don't even need to open a python interpreter or even know it's a python tool.

To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information
Assignee
Assign to
Time tracking