DCGAN for CelebA evaluated with FID (batched version)

DCGAN fork from https://github.com/carpedm20/DCGAN-tensorflow

Precalculated real world / trainng data statistics can be downloaded from: http://bioinf.jku.at/research/ttur/ttur.html

Usage

• Copy the file fid.py from TTUR root into the DCGAN_FID_batched directory
• Modify the dataset variable in run.sh
• Modify the data_path variable in run.sh
• Download the precalculated statistics (see above) and save them into the "stats" folder.
• Modify the incept_path in file run.sh
• Run the command: bash run.sh <disc. lr> <gen. lr>
• Checkpoint, sample and Tensorboard log directories will be automatically created in logs.

FID evaluation: parameters fid_n_samples and fid_sample_batchsize

The evaluation of the FID needs the comparison between precalculated statistics of real world data vs statistics of generated data. The calculation of the latter is a tradeoff between number of samples (the more the better) and available hardware. Two parameters in run.sh are concerned with this calculation: fid_n_samples and fid_sample_batchsize. The first parameter specifies the number of generated samples on which the statistics are calculated. Since this number should be high, it is very likely that it is not possible to generate this amount of samples at once. Thus the generation process is batched with batches of size fid_sample_batchsize.