[Question] How many images were used for calculating the metrics reported on the SFHQ dataset?

Hi,

Great work! I have a few questions regarding "Table 2: Quantitative results for face editing on SFHQ for ‘wearing glasses.’" Could you please provide more details on how to reproduce the results reported in this table? Specifically:
- How many images were used in the evaluation?
- Did the authors use the same prompt, "wearing glasses," for all images?

Thank you!