I am looking for a pre-trained model to extract features to use for trainsfer learining. The target is to train a linear regressive model that predicts facial blendshapes given an input face image. I use a vggface2 model from hugginface now, but I would like to know if NVIDIA has something similar available that I can use.
Hi @abraham.martin - I would recommend you start at the NVIDIA Tao Toolkit TAO Toolkit | NVIDIA Developer to see our Vision Transformer models. Best, Sophie