Hi, I was wondering whether the code for evaluation metrics is available? Specifically, I was wondering which repo did you use for Intra-Style-L? I tried with https://github.com/gordicaleksa/pytorch-neural-style-transfer/tree/a244dc61797f920a62146d98de92384352a4d3a2. The relative comparison with MD is still good but the scale seems to be different from the paper. E.g. I get 0.11 for MultiDiff and 0.088 for SyncDiff. (BTW I can match LPIPS score with the results I used for Intra-Style-L)
Hi, I was wondering whether the code for evaluation metrics is available? Specifically, I was wondering which repo did you use for Intra-Style-L? I tried with https://github.com/gordicaleksa/pytorch-neural-style-transfer/tree/a244dc61797f920a62146d98de92384352a4d3a2. The relative comparison with MD is still good but the scale seems to be different from the paper. E.g. I get 0.11 for MultiDiff and 0.088 for SyncDiff. (BTW I can match LPIPS score with the results I used for Intra-Style-L)