ElenaRyumina

/

face_emotion_recognition

like 63

Hi,

There was a typo on the main page. You need to run run_webcam which is located in "files and versions". Thanks, I have corrected it now.

Thank you for your quick answer!

I was also confused by the difference between the files and versions on Hugging Face and the files on the linked GitHub repo where I could not find the run_webcam.

Thank you very much for your interest in our work.

There are also similar files on GitHub: check_backbone_models_by_webcam and check_temporal_models_by_webcam. They work with torchscript models that do not require prior initialization of the model architecture, are designed for inference, and work faster than models uploaded to Huggingface. However, such models are not intended for fine-tuning. The files on GitHub have been created earlier. However, when creating a demo on Huggingface (https://huggingface.co/spaces/ElenaRyumina/Facial_Expression_Recognition), the use of torchscript models limited the possibilities of implementing the demo. So I uploaded standard torch models to Huggingface as well as another file for running models on a webcam.