search

dusty-nv / jetson-containers

Machine Learning Containers for NVIDIA Jetson and JetPack-L4T
MIT License
2.09k stars 435 forks source link

Track Anything Model (TAM) Example Broken #326

Open jimmycfa opened 9 months ago

jimmycfa commented 9 months ago

The TAM example gets an error on line 431 in app.py from TAM. Looks like gradio was not pinned in the original Docker and newer versions don't support the arguments on these two rows: https://github.com/gaomingqi/Track-Anything/blob/master/app.py#L431-L432

As opposed to fixing the app.py file I reverted to an older version of gradio in the kickoff of Docker:

./run.sh $(./autotag tam) sh -c "pip3 install gradio==3.38.0 ; cd /opt/Track-Anything ; python3 app.py"

I guess I could have also rebuilt the Docker image with gradio repinned to 3.38.0 too...

dusty-nv commented 9 months ago

Thanks @jimmycfa - pinned gradio in https://github.com/dusty-nv/jetson-containers/commit/3faa757eace8b0b25527671deadaa6db20eceeb5 and rebuilding the tam container 👍