Jia Guo
Jia Guo
@acakici2020 Done, please update to ver 0.6.2.
make sure you are using onnxruntime-gpu
@joytsay Yes and this 4.31ms includes post-processing.
You can now use https://github.com/deepinsight/insightface/blob/master/python-package/insightface/data/rec_builder.py Just call add() to add one identity with its images or add_image() to append one image with its label. All images should be already aligned.
please use the latest code
因为license原因没办法直接提供
Hi @SthPhoenix, thanks for your attention. 1. Just fixed. 2. For models that support dynamic input, you should pass the `input_size` param for detect() method. 3. It actually depends on...
BTW: if you're in a single-face situation, input size of 384/256(even 128) without padding is recommended.
My suggestion( if you want to use the pre-trained models ): Option 1: Use 512 input size Option 2: Use a combination of 640 and 256 inputs with some engineering...