caleboleary/MultiArc2Face

[Research]: Average multiple face embeddings of the same subject for improved results

1

stars

26

commits

Python

primary language

Jul 23, 2024

updated

README

Originally forked from this, thanks camenduru!

MultiArc2Face

This repo builds upon the work done to train the Arc2Face model by foivospar.

Specifically, this epxlores getting the embeddings of multiple of the same face and averaging them in different ways to gain more, and hopefully more correct, information about the face one is trying to reproduce with the Arc2Face model.

To run in google colab:

%cd /content
!GIT_LFS_SKIP_SMUDGE=1 git clone -b main https://github.com/caleboleary/MultiArc2Face
%cd /content/MultiArc2Face

!wget https://huggingface.co/camenduru/Arc2Face/resolve/main/freddie.png -O /content/MultiArc2Face/assets/examples/freddie.png
!wget https://huggingface.co/camenduru/Arc2Face/resolve/main/jackie.png -O /content/MultiArc2Face/assets/examples/jackie.png
!wget https://huggingface.co/camenduru/Arc2Face/resolve/main/joacquin.png -O /content/MultiArc2Face/assets/examples/joacquin.png
!wget https://huggingface.co/camenduru/Arc2Face/resolve/main/scrfd_10g_bnkps.onnx -O /content/MultiArc2Face/models/antelopev2/scrfd_10g_bnkps.onnx
!wget https://huggingface.co/camenduru/Arc2Face/resolve/main/arcface.onnx -O /content/MultiArc2Face/models/antelopev2/arcface.onnx

!pip install -r requirements.txt
!pip install gradio==4.25.0

!python app.py

Sample using median option, non-cherry picked

sample of different face counts

Contributors

caleboleary

21 commits

foivospar

4 commits

camenduru

1 commits

caleboleary/MultiArc2Face

[Research]: Average multiple face embeddings of the same subject for improved results

1

stars

26

commits

Python

primary language

Jul 23, 2024

updated

README

Originally forked from this, thanks camenduru!

MultiArc2Face

This repo builds upon the work done to train the Arc2Face model by foivospar.

Specifically, this epxlores getting the embeddings of multiple of the same face and averaging them in different ways to gain more, and hopefully more correct, information about the face one is trying to reproduce with the Arc2Face model.

To run in google colab:

%cd /content
!GIT_LFS_SKIP_SMUDGE=1 git clone -b main https://github.com/caleboleary/MultiArc2Face
%cd /content/MultiArc2Face

!wget https://huggingface.co/camenduru/Arc2Face/resolve/main/freddie.png -O /content/MultiArc2Face/assets/examples/freddie.png
!wget https://huggingface.co/camenduru/Arc2Face/resolve/main/jackie.png -O /content/MultiArc2Face/assets/examples/jackie.png
!wget https://huggingface.co/camenduru/Arc2Face/resolve/main/joacquin.png -O /content/MultiArc2Face/assets/examples/joacquin.png
!wget https://huggingface.co/camenduru/Arc2Face/resolve/main/scrfd_10g_bnkps.onnx -O /content/MultiArc2Face/models/antelopev2/scrfd_10g_bnkps.onnx
!wget https://huggingface.co/camenduru/Arc2Face/resolve/main/arcface.onnx -O /content/MultiArc2Face/models/antelopev2/arcface.onnx

!pip install -r requirements.txt
!pip install gradio==4.25.0

!python app.py

Sample using median option, non-cherry picked

sample of different face counts

Contributors

caleboleary

21 commits

foivospar

4 commits

camenduru

1 commits

Languages

Python

100.0%