/Text-2-Image-Flowers

Text to Image synthesis for Oxford Flowers Dataset

Primary LanguagePythonMIT LicenseMIT

Text 2 Image Flowers

  1. Set up a virtual environment
python3 -m venv venv
  1. Configure enevironment using the following command:
pip install -r requirements.txt
  1. Run the following command to train the model:
python runtime.py --cfg config.yml

  1. Open config_test.yml and change the TEST key to appropriate value:

    • 1 for test on the validation set
    • 2 for test on predifined cutsom captions
    • 3 for generating model summary and architecture

  2. Run the following command to test the model:

python runtime.py --cfg config_test.yml

Note: A checpoint is avaible at https://drive.google.com/file/d/1-b1MiUPppgY9zeNQCBqvhQIJRZV0BHBa/view?usp=sharing

Credits:

  1. Recurrent-Affine-Transformation-for-Text-to-image-Synthesis
  2. RAT-GAN