Program Arguments

There are a few arguments we can change that can influence the model, environment or debugging of this notebook.

Environments: For all possible environments please have a look at: OpenAI Environments Do note however that I have not tested this behaviour and you'll likely have to tune your hyperparameters accordingly and change some code

Variable	Description	Default	Possible values	Type
	ENVIRONMENT OPTIONS
ENVIRONMENT	The selected environment (from the environments list or just typed in by hand)	'CartPole-v1'	A string representing a gym environment	str
	DEBUG OPTIONS
VERBOSE	What types of debug data are printed. 0 for no debug printing, 1 for normal debug printing and 2 for normal + GPU device placement printing (this option is very spammy and resource intensive)	1	0/1/2	int
VISUALIZE['train']	Whether to visualize training	False	-	bool
VISUALIZE['evaluate']	Whether to visualize evaluation	True	-	bool
	AGENT OPTIONS
MODE	What the program should be doing	'both'	'train' / 'evaluate' / 'both'	str
SCORE_H5	If you already have a saved model load it from ./{ENVIRONMENT}-{SCORE_H5}.h5	None	-	str
TRAIN_STEPS	The amount of training games we wish to play	500	> 0	int