Test data

[mishimario]

Hi, do you have a plan to release the test data of Synth-500 and AgeDB? Thank you in advance!

[livermoressss]

Also want to know the evaluation detail of FID. What would be number of dim for the Inception V3?

[mishimario]

And I want to know about preprocessing when evaluating. In the paper, there is face detection and alignment. What is this alignment? It's not only removing the background?

[AvirupJU]

And I want to know about preprocessing when evaluating. In the paper, there is face detection and alignment. What is this alignment? It's not only removing the background?

The alignment is based on landmarks. Check out the run_emoca() method in gradio_demo/app_controlnet.py

[mishimario]

Thank you. I got it. How about the number of dim for FID? The dim is 2048?

[foivospar]

Hi, do you have a plan to release the test data of Synth-500 and AgeDB? Thank you in advance!

Please see here for test images.

Thank you. I got it. How about the number of dim for FID? The dim is 2048?

Given the number of test images, we used dim=192 for FID.