Implementation of Generating Videos with Scene Dynamics.
You can star this repository to keep track of the project if it's helpful for you, thank you for your support.

for example, you can download:
<Golf Only> → http://data.csail.mit.edu/videogan/golf.tar.bz2
<List of Golf Files> → http://data.csail.mit.edu/videogan/golf.txt

for example, if you have downloaded the data of golf, then you can execute:
"bunzip2 golf.tar.bz2"
"tar xvf golf.tar"

options explain:
-info: 
	Introduce options.
-batch_size: 
	Number of videos to process in a batch.
-lr_g: 
	Learning rate for generator.
-lr_d: 
	Learning rate for discriminator.
-beta1_g: 
	Beta1 for generator.
-beta1_d: 
	Beta1 for discriminator.
-dis_dim: 
	Scale for discriminator-channel.
-gen_dim: 
	Scale for generator-channel.
-gen_scale: 
	Scale for generator-width and height.
-pic_dim: 
	Dimension of image color.
-noise_dim: 
	Dimension of initial noise vector.
-sample_size: 
	Number of samples to be generated at once(for evaluate).
-mask_L1_lambda: 
	Weight for L1 regularizer of mask.
-trainlogfile: 
	Record the train info(filename).
-modelSaved: 
	Save the trained model(path).
-samplesSaved: 
	Save the sample videos from generator(path).
-max_epoch: 
	Number of training epochs.
-save_interval: 
	Save and test the model each save_interval epochs.
-trainSet: 
	The paths of the videos for training.
-imgSize: 
	The size of img(each frame) in the videos(trainSet).

make sure that you have installed all of the dependencies mentioned in Dependency.
you can run "pip install -r requirements.txt" to install these dependencies.