SthPhoenix

Buffalo_l consists of `scrfd_10g_bnkps` detection and `w600k_r50` recognition models. You can get both models from Google drive links located at my [repo](https://github.com/SthPhoenix/InsightFace-REST#list-of-supported-models) `w600k_r50` model is the same as in package,...

1. That's great! Thanks! 2. Yes, I've just meant that example code will throw exception out of box, though easily fixable. 3. New detector works just great for other image...

Thanks! I was investigating these options yesterday, option 2 is more promising but more logically complicated. In a long term retraining seems a better solution, could you give any hints...

Little update: this bug seems to be related only to `*_bnkps` models. When using models without key points detection work as expected

@nttstar , I have retrained `scrfd_2.5g_bnkps`, with batch norm replaced with group norm (new model should be called `scrfd_2.5g_gnkps` i think), just like for `scrfd_2.5g` model, I achieved following WiderFace...

> @SthPhoenix Thanks! Did you make the feature maps shared? No, shared feature maps seems to reduce accuracy more noticably. >BTW, you can open a new repo to place this...

> @SthPhoenix So you still using widerface, only change the config? Yes, just it.

I have tested it by modifying `scrfd_500m.py `(as it could be faster trained) as follows: ``` norm_cfg=dict(type='GN', num_groups=16, requires_grad=True), cls_reg_share=True, strides_share=True, ``` After 640 epoch I've got following mAP: `0.881,...

> Shared feature map should be better by using GN, from my experiments of resnet based backbone. Hmmm, I'll check it on other models, thanks!

I have released retrained models at my repo. Models accuracy on WiderFace benchmark: | Model | Easy | Medium | Hard | | -------------------| -------|----------| ----- | | scrfd_10g_gnkps |...