How to use the best model in dynamic v1.0.0

[Penguin-is-rich]

Hello Authors,

I hope this message finds you well. I'm currently attempting to utilize your latest released best model, but even after reading the README, I'm still having trouble understanding it thoroughly.

I would like to inquire about the following:

1. Where should the "r2plus1d_18_32_2_pretrained.pt" and "deeplabv3_resnet50_random.pt" file be placed within the directory structure for the program to function properly?

2. Regarding the instructions in the README, are they intended solely for testing purposes, or do they involve retraining? The output I've observed seems to resemble a retraining process.

3. It would be greatly appreciated if there were more detailed instructions available.

Thank you for taking the time to respond.

Best regards