邸德宁

You use 0.0001 as base learning rate? Or you means when your lr up to 0.0001, the loss get larger? Can use share more details about this appearance? Maybe some...

Read the function named *mx2tfrecords* at ./data/mx2tfrecords/, and '__main__' part in the same file shows its invoking process. You need to download the origin dataset (see: https://github.com/deepinsight/insightface), then try to...

@muneebullahkhan I guess you are using ipython. Maybe you need to check whether the path 'image/test.jpg' is consistent with your workdir. At least， I don't think there is any mistake...

no, of course not. In fact, this dist represents cos distance. dist = 2 * ( 1 - cos ).