Jia Guo
Jia Guo
@zerodwide SCRFD-10GF @javiabellan You should do face-alignment, not image resizing.
No such plan. For r100-w2m you can train it by yourself.
You should generate them by yourself.
the version of onnxruntime-gpu library should match your cuda version.
cuda>=10.2 and onnxruntime-gpu>=1.6 is recommended.
@dhagarwa try installing the latest version, we have put numpy into the build system requirements.
bad news, I will check the url soon
The file server donated by my friend is down. We will try to find an alternative server, but not sure. Before that, we will upload the model files to google...
@sanidhyaagrawal @gregordr Hi please download the following manual model packages and unzip them under ``~/.insightface/models/`` buffalo_l: https://drive.google.com/file/d/1qXsQJ8ZT42_xSmWIYy85IcidpiZudOCB/view?usp=sharing buffalo_m: https://drive.google.com/file/d/1net68yNxF33NNV6WP7k56FS6V53tq-64/view?usp=sharing buffalo_s: https://drive.google.com/file/d/1pKIusApEfoHKDjeBTXYB3yOQ0EtTonNE/view?usp=sharing buffalo_sc: https://drive.google.com/file/d/19I-MZdctYKmVf3nu5Da3HS6KH5LBfdzG/view?usp=sharing antelopev2: https://drive.google.com/file/d/18wEUfMNohBJ4K3Ly5wpTejPfDzp-8fI8/view?usp=sharing
@sanazkhalili pip 0.6 is updated which supports model downloading again, you may check.