SairamYadla/VGG_Facial_Emotion_Detection

Facial Emotions Detection VGG

This project was a part of my LMU Computer Vision and Deep Learning internship for facial emotions classifications.

Recreated a simpler VGG version and trained on small fer dataset to learn the Hugging Face process of uploading and testing model with Gradio

This model is a VGG-based convolutional neural network built in PyTorch for facial emotion recognition.
It classifies faces into one of 7 emotions : Angry, Disgust, Fear, Happy, Neutral, Sad, and Surprise by using images resized to 64×64 RGB.

The network was trained on the FER-2013 dataset from Kaggle (msambare/fer2013), which contains labeled facial images.

metrics Epoch 39: train loss 0.1205, acc 0.9637, val loss 2.1266, acc 0.6254

Author: Sairam Yadla
License: MIT
Hugging Face Repository: SairamYadla/Facial_Emotions_Detection

Github Repository: SairamYadla/Facial_Emotions_Detection