InsightFace REST API for easy deployment of face recognition services with TensorRT in Docker.
Apache License 2.0
488
stars
118
forks
source link
Thankyou for excllent work. I am able to run demo_client but please i want to know how i can select the model of my choice or put custom trained model. #122
To select different models from supported list you should just replace following lines in compose/.env:
DET_NAME=yolov5l-face
REC_NAME=w600k_r50
Adding custom detector is a bit tricky since you need to implement postprocessing specific for model, while face recognition models based on official InsightFace arcface architecture can be just placed in folder models/onnx/custom_rec_model.
To select different models from supported list you should just replace following lines in
compose/.env
:Adding custom detector is a bit tricky since you need to implement postprocessing specific for model, while face recognition models based on official InsightFace arcface architecture can be just placed in folder
models/onnx/custom_rec_model
.If you need to add additional compatible rec models they should be declared in https://github.com/SthPhoenix/InsightFace-REST/blob/master/src/api_trt/modules/configs.py You can use
custom_rec_model
config as reference