lymanblue

Question about is_bgr flag

5

Hi~ We haved tried the following args for preprocessing_args, but the converted coreml model still requires BGRA input. args = dict(is_bgr=False, red_bias=-1, green_bias=-1, blue_bias=-1, image_scale=2/255.0) Are there any other settings...

Should we apply face alignment for evaluation data?

11

Hi~ Thank you for your great work. Does the reported accuracy result on validation data (e.g. LFW, MegaFace) apply face alignment process (e.g., MTCNN)? Thank you.

Training epochs for MobileFaceNet

1

Hi~ We trained mobilefacenet with 60 epochs on CASIA (small protocol) but still cannot achieved the reported accuracy (before removing noise). Initial LR=0.1, and divided by 10 at epoch 36,...

Questions about the performance of megaface with @R notation

7

Hi~ Is the symbol @R represent the cleaned megaface dataset as stated in insighface paper? The performance is improved significantly with symbol R in your experiments. We tried the cleaned...

Question about training epochs

1

Hi~ Are the results reproducible from the default settings (e.g., total epochs = 11) in train.py? Are there different hyper-parameters for different CNN under different scenarios? (e.g., ResNet vs. MobileFaceNet,...

How evaluate the performance on MegaFace?

10

Hi~ Thank you for your great work! The megaface devkit requires 2 files (i.e., megaface and facescrub) for performance evaluation. However, the eval_megaface.py generates .feat in each folder. (do we...