-
Notifications
You must be signed in to change notification settings - Fork 26
Description
Thank you for your impressive work presented in your paper. We greatly appreciate the detailed benchmarking framework you established, particularly the comprehensive evaluation metrics for Reference Consistency (including facial, clothing, object, and background consistency), Instruction Following, and Visual Quality.
However, we noticed that while the evaluation methodology is thoroughly described, the actual implementation or code for the Reference Consistency assessment module has not been open-sourced. This makes it difficult for researchers and developers in the community to reproduce your evaluation protocol or fairly compare their own models against your benchmarks using the same criteria.
Would it be possible to open-source the Reference Consistency evaluation code (e.g., the scripts or models used to measure facial/clothing/object/background consistency)?
Thank you very much for your time and consideration. We look forward to your response.
Best regards,