btgym.monitor package

btgym.monitor.tensorboard module

class btgym.monitor.tensorboard.BTgymMonitor(scalars={}, images={}, histograms={}, text={}, logdir='./tb_log', subdir='/', purge_previous=True, **kwargs)[source]

Light tensorflow ‘summaries’ wrapper for convenient tensorboard logging.

Monitor parameters: Sets of names for every value category: scalars, images, histograms ant text. logdir - tensorboard log directory; subdir - this monitor log subdirectory; port - localhost webpage addr to look at; reload - web page refresh rate. purge_previous - delete previous logs in logdir/subdir if found.

write(feed_dict, global_step)[source]

Updates monitor with provided data.

class btgym.monitor.tensorboard.Tensorboard(logdir='./btgym_log', port=6006, reload=30)[source]

Utility class to start/stop tensorboard server.

start()[source]

Launches Tensorboard app.

stop()[source]

Closes tensorboard server.

btgym.monitor.tensorboard2 module

class btgym.monitor.tensorboard2.BTgymMonitor2(scalars={}, images={}, histograms={}, text={}, logdir='./tb_log', subdir='/', purge_previous=True, **kwargs)[source]

Light tensorflow ‘summaries’ wrapper for convenient tensorboard logging.

Monitor parameters: Sets of names for every value category: scalars, images, histograms ant text. logdir - tensorboard log directory; subdir - this monitor log subdirectory; port - localhost webpage addr to look at; reload - web page refresh rate. purge_previous - delete previous logs in logdir/subdir if found.

write(sess, feed_dict, global_step)[source]

Updates monitor with provided data.

class btgym.monitor.tensorboard2.Tensorboard(logdir='./btgym_log', port=6006, reload=30)[source]

Utility class to start/stop tensorboard server.

start()[source]

Launches Tensorboard app.

stop()[source]

Closes tensorboard server.