Why does it take a long time to run a project on Docker?

[1ADZX]

downloading: "http://cmp.felk.cvut.cz/~mishkdmy/models/loftr_outdoor.ckpt" to /root/.cache/torch/hub/checkpoints/loftr_outdoor.ckpt
100%|██████████████████████████████████████| 44.2M/44.2M [24:00<00:00, 32.2kB/s]
[2024/08/23 12:22:44 hloc INFO] Loaded LoFTR with weights outdoor
[2024/08/23 12:22:44 hloc INFO] Loading model using: 1446.551s
[2024/08/23 12:22:49 hloc INFO] Matching images done using: 5.529s
[2024/08/23 12:22:50 hloc INFO] ransac_method: CV2_USAC_MAGSAC, geometry_type: Fundamental
[2024/08/23 12:22:50 hloc INFO] ransac_method: CV2_USAC_MAGSAC, geometry_type: Homography
[2024/08/23 12:22:50 hloc INFO] RANSAC matches done using: 0.617s
[2024/08/23 12:22:50 hloc INFO] Display matches done using: 0.360s
[2024/08/23 12:22:50 hloc INFO] TOTAL time: 1453.248s
[2024/08/23 12:22:50 hloc INFO] Dump results done!

My graphics card model is GeForce RTX 3080 12GB，The specific parameters for matching the two images are shown in the figure。

[1ADZX]

The specific parameters for matching the two images are shown in the figure.

[1ADZX]

I think it takes too long to download the model.

[1ADZX]

Can you design a method to load the model locally when using Docker? It is too slow to download the model again every time.