Open ZhenyuSun-Walker opened 2 months ago
Thank you for your inquiry. We directly used the metrics from the original paper and ensured that our testing environment was consistent with that of the paper to maintain comparability of results.
Thank you for your suggestion! However, I find it strange when calculating the text2room rendering metrics. How do you get the render images and the gt images?
I found that the rendered image in text2room is quite low-quality. How can I correctly calculate the metrics?
Dear sir, I really appreciate your outstanding work, but would you mind guiding me how to properly calculate the metric you measured in the paper, when using LucidDreamer, as well as in your work? I find it strange to calculate the correct metric, especially when coping with the CLIP Score and the CLIP IQA.