Language GANs Falling Short

Code for reproducing all results in our paper, which can be found

(key) Requirements

• Python 3.x
• Pytorch 0.4.1
• TensorboardX

Structure

• common folder: most of the important code is here, including all models and utilities
• synthetic_data_experiments folder: code to run and reproduce all oracle experiments
• real_data_experiments folder: code to run results for ImageCoco and News datasets

Reproducibility

• For synthetic data, simply run oracle_eval.py found in the synthetic_data_experiments folder.
• For real data, we uploaded the weights (and corresponding hyperparameters) in real_data_experiments/trained_models folder. You can load the model by using the --load_{gen/disc}_from_file argument. For example,

python main.py --load_gen_path trained_models/news/word/best_mle

Contact

For any questions / comments / concerns, feel free to open an issue on github, or to send me an email at
my_email_address.

We strongly believe in fully reproducible research. To that end, if you find any discrepancy between our code and the paper, please let us know, and we will make sure to address it.

Happy text modeling :)