/cycle-gan-implementation

Primary LanguageJupyter Notebook

Проверка:

Алексей уже дал добро, мой тг: https://t.me/andregrit

Monet style telegram bot

Based on CycleGAN architecture Source

Alt text

Steps

  1. Ensure that the GPU driver is installed properly on the instance: CUDA installation guide.
  2. Create .env file according to example.env
  3. Use docker-compose to build the backend.
docker-compose --profile prod up

Training

In order to train the model

docker-compose --profile dev up

connect to container (or use .devcontainer in vscode), and see notebooks with examples.

To display training graphs, the Visdom server was configured, which will be available at http://localhost:8097.