Closed Varun-GP closed 6 months ago
Hi @Varun-GP, I uploaded a Script to use a trained model (e.g. AffectNet8Combined) with your webcam, called "inference_on_webcam.py". If you want to use the model on videos/images, just apply the model directly to an image or to each video frame instead of passing the webcam to it.
Keep in mind to change the model-path, your torch-device (cuda, mps, ...) and the number of output neurons based on the used model (e.g. the Affectnet8Combined model has 10 output neurons, 8 for the discrete classes Angry, Sad etc. and two neurons for Valence and Arousal).
Best regards, Niklas
Thanks a ton!
Thanks a lot for this wonderful work! I wanted to test the model on new videos/images. Can you please provide any live video expression inference?