def collect_initial_episodes(config):
if config.source_train_dir:
tools.copy_directory(config.source_train_dir, config.train_dir,
config.source_train_amount)
if config.source_test_dir:
tools.copy_directory(config.source_test_dir, config.test_dir,
config.source_test_amount)
items = config.random_collects.items()
items = sorted(items, key=lambda x: x[0])
for name, params in items:
message = 'Collecting {}+ random episodes ({}).'
tf.logging.info(message.format(params.num_episodes, name))
control.random_episodes(params.task.env_ctor, params.num_episodes,
params.save_episode_dir)
def collect_initial_episodes(config):
items = config.random_collects.items()
items = sorted(items, key=lambda x: x[0])
existing = {}
for name, params in items:
outdir = params.save_episode_dir
tf.gfile.MakeDirs(outdir)
if outdir not in existing:
existing[outdir] = len(tf.gfile.Glob(os.path.join(outdir, '*.npz')))
if params.num_episodes <= existing[outdir]:
existing[outdir] -= params.num_episodes
else:
remaining = params.num_episodes - existing[outdir]
existing[outdir] = 0
env_ctor = params.task.env_ctor
print('Collecting {} initial episodes ({}).'.format(remaining, name))
control.random_episodes(env_ctor, remaining, outdir)