The Repository for the baselines for GanDraw Dataset
We Visualize the Training Progress with Visdom. The visdom port for each server is listed as following:
- nlp.cs.ucdavis.edu: 8097
- interaction.cs.ucdavis.edu: 8098
To switch between using a dataset with image dimension of 128 x 128 and a dataset with image dimensino of 64 x 64, you need to change the following in gandraw_args.json:
- img_size: 128 -> 64
- dataset: gandraw -> gandraw_64
- test_dataset: gandraw_test -> gandraw_64_test
- val_dataset: gandraw_val -> gandraw_64_val
- gan_type: recurrent_gan_mingyang -> recurrent_gan_mingyang_img64
The expriment and results_path should also be changed accordingly.
- Move the GeNeVa Baselines to This Repository.
- Create the Readme on preprocessing GANDRAW dataset.
- Creat the Readme on running Baseline1
- Creat the Readme on running Teller
- Create the Baselines for Drawer
- Create the Readme on Drawer
- Create the Code for Baseline1 Evaluation
- Create the Code for Teller Evaluation.
- Create the Code for Drawer Evaluation.