Style2Vec

Before exploring this project, ensure that all required dependencies are installed:

pip install -r requirements.txt

Training

After preparing your own dataset, you can train the model using one of the following commands:

• Basic Training:

  python train_style2vec.py

  This command initiates the training process using the basic Style2Vec model on your dataset.

• Cosine Similarity Training:

  python train_style2vec_dot.py

  Use this command to train the model with cosine similarity, which is useful for certain types of similarity searches.

Inference

You can run the following command to obtain the embedding of a handwritten sample:

python inference.py path/to/sample1.png path/to/sample2.png --model_path path/to/model

Demo

From the main folder, you can run one of the following commands to see the demo in action:

• Basic Prediction Demo (with human-readable probabilities):

  python match_prediction_demo.py sample/test/font_0_sample_0.png --candidate_file demo/demo_candidates.txt --visualize --output demo/demo.png --prob_format human

  This command runs a demo that computes predictions using the basic model and displays probabilities in a human-friendly format.

• Cosine Similarity Prediction Demo:

  python match_prediction_demo_dot.py sample/test/font_0_sample_0.png --candidate_file demo/demo_candidates.txt --visualize --output demo/demo.png --prob_format none

  This variant uses cosine similarity for prediction, outputting results without human-readable probability formats.

• Heatmap Demo:

  python heatmap_prediction_demo_dot.py --candidate_file demo/demo_candidates_heatmap.txt --output demo/demo_heatmap.png

  With this command, the demo generates a heatmap visualization based on the candidate file, illustrating the prediction probability distribution.