use host directory to save CLIP model for nanoowl

dusty-nv / jetson-containers

Machine Learning Containers for NVIDIA Jetson and JetPack-L4T

MIT License

1.89k stars 416 forks source link

use host directory to save CLIP model for nanoowl #504

Closed TadayukiOkada closed 1 month ago

TadayukiOkada commented 1 month ago

fix #454 not sure if this is the right fix, but I linked the /root/.cache/clip to /data/models/clip inside the container so that the CLIP models downloaded will be saved in the host directory. Also, updated README.md to reflect jetson-containers command addition.

dusty-nv commented 1 month ago

Thanks @TadayukiOkada, appreciate it :) merged this and rebuilding the nanoowl containers now.