A model merge of 3 small model based on Efficient Net B4, Efficient Net B0 and SR CNN architecture using pytorch. Here, i use 2 custom dataset: gender and age dataset and emotion dataset .

I. Build model for Gender and Age recognition task by a Multi-task model: Result

II. Build model for Emotion recognition task: Result

III. Merge them and use Mediapipe, Opencv library for recognition

• Mediapipe
• Opencv
• Numpy
• Python3
• Pytorch

pip install mediapipe    # mediapipe library
pip install opencv       # opencv library
pip instal numpy         # numpy library
pip instal pytorch       # pytorch library
pip install torchsummary # summary
pip install torchvision  # pytorch for vision

NB: Update the libraries to their latest versions before training.

⬇️⬇️Download and extract all my train dataset on Kaggle: Gender and Age Dataset

⬇️⬇️Download pretrained model: Model

Run the following scripts for training and/or testing

python train.py # For training the model 

Run the following scripts for visual result of model:

1. Download Docker

Open CMD

2. Download my image

docker pull vvduc1803/gender_age_emotion:latest                                  # Pull image

3. Copy and paste

docker run -it -d --name gender_age_emotion vvduc1803/gender_age_emotion  # Run container

4. Copy and paste

docker run gender_age_emotion                                             # Run visual result

Recognition results

1. Experiments with different learning-rate and optimizers.
2. Converting and optimizing pytorch models for mobile deployment.

Van Duc