README.md

Hairstyle Recommendation App 2023.11

About

Hairstyle Recommendation App is a separate service which takes image of client, applies pipeline of recommandation algorithms and returns recommendation scores for each hairstyle.

Installation

• Conda (https://docs.conda.io/en/latest/miniconda.html)
• Tensorflow (https://www.tensorflow.org/install/pip)
• Python packages
  • pip install tensorflow-gpu
  • pip install pillow opencv-python numpy
• (optional) Nvidia GPU support
  • CUDA 11.2 (https://developer.nvidia.com/cuda-toolkit-archive)
  • cuDNN 8.1 (https://developer.nvidia.com/rdp/cudnn-download)
• Flask: pip install flask flask-cors

Docker (recommended)

• Docker (https://docs.docker.com/engine/install/ubuntu/)
• (optional) nvidia-docker2 (https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/install-guide.html)
• tensorflow for Docker (https://www.tensorflow.org/install/docker)
• Build dev-server application: DOCKER_BUILDKIT=1 docker build -t hairstyler-dev -f Dockerfile.dev .
  • (optional) Build dev-server-application with GPU support: DOCKER_BUILDKIT=1 docker build -t hairstyler-gpu-dev -f Dockerfile.gpu.dev .
• Build tests: DOCKER_BUILDKIT=1 docker build -t hairstyler-tests -f Dockerfile.tests .
  • (optional) Build tests with GPU support: DOCKER_BUILDKIT=1 docker build -t hairstyler-gpu-tests -f Dockerfile.gpu.tests .

Running

To run dev-server application execute following command: flask --app hairstyler run --cert=certificates/cert.pem --key=certificates/key.pem --host=0.0.0.0.
To run unit tests execute: python -m unittest discover -v.

Docker (recommended)

• To run dev-server application execute following command: docker run -p 5000:5000 hairstyler-dev
  • (optional) To run dev-server with GPU support execute: docker run --gpus all -p 5000:5000 hairstyler-gpu-dev
• To run unit tests execute: docker run hairstyler-tests
  • (optional) To run unit tests with GPU support execute: docker run --gpus all hairstyler-gpu-tests