icAN

Style transfer is an important area of research in the field of generative AI, with applications in a wide range of fields, including art, design, and advertising. This work explores the use of GAN in style transfer for icons across different platform styles.

Code Structure

The project structure is as follows:

• ckpt: Directory to store the trained model checkpoints.

  • model_demo.pt: Trained model checkpoint for demo.

• dataset.py: Python script for handling the dataset.

• data/: Directory containing the training.

  • raw: Raw RGB images for icons.
  • edge: Edges for icons.
  • meta: Information about icons including reference icons in dataset.

• main.py: Main Python script for running the project.

• models/: Directory containing models for this project.

  • basic: The baseline model for this project.
  • resnet: The model using ResNet structure.

• requirements.txt: File listing the project's dependencies.

• scrawler/ : Directory containing utility functions for data preparation.

  • get_icon_names.py: A script that can automate the retrieval of icon names.
  • download.py: A script that can download data from https://icons8.com according to the given icon names list.
  • download_multi.py: A script that can download data in multi-threading mode.
  • preprocess.py: A script that can preprocess downloading data.

• eval: Directory containing evaluation during training

Usage

1. Ensure that you have the required dependencies installed by running: pip install -r requirements.txt

2. Interact with the trained model using demo.ipynb.

3. Run main.py to train a new model.