Brad Dwyer
Brad Dwyer
It's not just `huggingface`; other libraries like `openai/clip` and `torch` also store data in `~/.cache`. Ideally make a way to mount volumes in custom locations where data will persist between...
I'd like to add my +1. Our lawyer has advised that the current case-law requires the user to actively check a box agreeing to the terms. The current implementation is...
This should be fixed in datasets downloaded with the latest version of `roboflow`. We’re now dynamically rewriting the yaml paths to the new format when `ultralytics>=0.30.0` is detected.
Could you link to the dataset on Roboflow? I can have a look and see if it looks like it's on the data export end or the model ingestion/dataloader end.
Are you using GroundingDINO or GroundedSAM?
Multi-GPU inference sounds cool if you want to submit a PR for parallelism. Though I’m not sure all the BaseModels would play nicely with that.
PR is updated in response to your comments here & @capjamesg's comments on Slack.
> Is tracker_length still needed? I'm using it in some client code to create animated `Annotator`s. The frame of the animation for a detection needs to be `tracker_length%num_frames`. Also planning...
Updated to remove `tracker_length` and `track_starts`.
Should we make the default be False? When would you want to overwrite it? (Exported versions are immutable)