Description:
Hello,
I am currently trying to evaluate face recognition models on the standard benchmark datasets (LFW, CFP, AGEDB) as per the instructions in your repository under eval/eval_recognition/. However, I am unable to access or download the required evaluation data from the provided links:
Google Drive Link (Code: z7hs)
Baidu Drive Link
Both links appear to be inaccessible—either the files are missing, the permission is denied, or the links have expired. Since these datasets are critical for running eval.sh and replicating your evaluation results (e.g., ./eval.sh magface_epoch_00025.pth official 100), I would appreciate it if you could do one of the following:
-
Provide updated and publicly accessible download links for the datasets.
-
Include a backup source or alternative method to obtain the required data.
-
Add the dataset as a downloadable release asset or support script (e.g., download_data.py) to fetch and prepare the datasets automatically.