This is the repository for Smart-Filter, a deep learning open source web app. This repository provides end-to-end pipeline from model architecture of face detection using transfer learning to deployment of a web application using Docker container.
Visit the official repository webpage: https://karan469.github.io/Smart-Filter/
Web app soon to be deployed.
- tensorflow-cpu==2.2
- pytorch==1.6.0+cpu: pip install torch==1.6.0+cpu torchvision==0.7.0+cpu -f https://download.pytorch.org/whl/torch_stable.html
- detectron2-cpu (for inference): python -m pip install detectron2 -f https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.6/index.html
- detectron2-gpu (for training): python -m pip install detectron2 -f https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.6/index.html
- CUDA (10.1)
Caution: Specified softwares must be installed in their correct versions. Mentioned versions are compatible with each other.
Deployment/
Web application for deploying a trained set of models on Flask using Docker containers. To read more about Docker, visit official Docker documentation: https://docs.docker.com/
- templates/
- uploads/
- app.py
- detectron.py
- smile.py
- utils.py
- facedetector.py
- requirements.txt
training/
Contains training modules for semantic segmentation, face detection and smile detection. iPython notebooks contains relevant model architectures.
- Face Detection using Detectron2/
- Final Ensemble/
- Key Points Detection/
- Semantic Segmentation/
- Smile Detection
results/
Contains demo images as examples of working prototype of final web application consisting of features such as custom background and caption writing according to facial features such as smile.
For more information visit the official documentations of different frameworks.
For any queries, email me at mailto://tkaran.iitd@gmail.com