Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Configurable Object Detection Model #89

Open
PaulHax opened this issue Aug 7, 2024 · 0 comments
Open

Configurable Object Detection Model #89

PaulHax opened this issue Aug 7, 2024 · 0 comments

Comments

@PaulHax
Copy link
Collaborator

PaulHax commented Aug 7, 2024

Short term motivation

Demo general usefullness of app by allowing user to quickly configure wide variarty of dataset+model combos.

Could have a GUI to select a hugging face COCO dataset, then all related models are shown in model dropdown. Or vice versa, first select model, then related datasets are shown.

Where do the models come from?

How to configure models?

  • CLI, like how we do COCO datasets. I.e. nrtk-explorer --models ../my-model/some-file
  • GUI to enter model name or pick file/directory

Need to convert model object detection output to COCO annotation data structure?

Code touch points

Embeddings too?

Shortest putt

Define list of Hugging Face model names via CLI.

Alternative

Support for visualization of offline model predictions (e.g. also loaded in COCO format), which allows for faster analysis and does not require on-device model inference.

  • Support for multiple models with different labels (e.g. “fasterrcnn”, “yolo”, “ground truth”)

Related

Expanding set of datasets/perturbations/models available, switching to JATIC protocols

A PR: #117

@PaulHax PaulHax changed the title Configurable Object Detection Model new feature Configurable Object Detection Model Sep 12, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant