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

Support uploading and selection of custom user models #1354

Open
Alextopher opened this issue Jun 28, 2024 · 4 comments
Open

Support uploading and selection of custom user models #1354

Alextopher opened this issue Jun 28, 2024 · 4 comments
Labels
enhancement New feature or request

Comments

@Alextopher
Copy link
Contributor

Alextopher commented Jun 28, 2024

Is your feature request related to a problem? Please describe.
The existing workflow for using custom object detection models isn't well suited to the task of experimenting with custom models or sharing resources across teams. Being required to sign into the coprocessor is an unnecessary barrier for team's who are just interested in downloading & using someone else's model.

Describe the solution you'd like
Add support for uploading custom models and a way to switch between models.

Describe alternatives you've considered
Module selection alone would be a good start. Teams could continue to use ssh to upload and manage models but then have access to an interface to pick-and-choose the model of interest.

Additional context

@Alextopher Alextopher added the enhancement New feature or request label Jun 28, 2024
@mcm001
Copy link
Contributor

mcm001 commented Jun 28, 2024

I think the simplest, MVP implementation of this feature is a CRUD interface in our settings tab to upload arbitrary files, and a dropdown in neural network pipelines to select what file name I want to use. This covers 90% of what users actually need I think.

@Alextopher
Copy link
Contributor Author

Sounds good! Tomorrow I should have the time to gauge if I'm able to make that contribution.

@Alextopher
Copy link
Contributor Author

Alextopher commented Jun 30, 2024

I'll take this on - but no promises on the timeline. I'm going to wait until the dependency/build issues are fixed

@mcm001
Copy link
Contributor

mcm001 commented Jun 30, 2024

Should (?) be fixed! Note my comment for windows users needing to use the FRCK JDK

mcm001 pushed a commit that referenced this issue Sep 21, 2024
This PR is for part 1 of #1354. It focuses on adding a model selection
interface for models that exist in `photonvision_config/models/`. Upon
completion we can ship more than 1 model and users could upload their
own through `ssh` without deleting the shipped model. This PR also adds
the abstractions need to support more DNN backends (say OpenCV, or RPI
AI Kit)

Up next is adding a CRUD interface for managing models through the UI.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
enhancement New feature or request
Projects
None yet
Development

No branches or pull requests

2 participants