gryf/coach
1
0
Fork 0
mirror of https://github.com/gryf/coach.git synced 2025-12-17 19:20:19 +01:00
Code Issues Projects Releases Wiki Activity

Cleanup imports.

Browse Source 
Till now, most of the modules were importing all of the module objects
(variables, classes, functions, other imports) into module namespace,
which potentially could (and was) cause of unintentional use of class or
methods, which was indirect imported.

With this patch, all the star imports were substituted with top-level
module, which provides desired class or function.

Besides, all imports where sorted (where possible) in a way pep8[1]
suggests - first are imports from standard library, than goes third
party imports (like numpy, tensorflow etc) and finally coach modules.
All of those sections are separated by one empty line.

[1] https://www.python.org/dev/peps/pep-0008/#imports
This commit is contained in:
Roman Dobosz
2018-04-12 19:46:32 +02:00
parent cafa152382
commit 1b095aeeca
75 changed files with 1169 additions and 1139 deletions
Download Patch File Download Diff File Expand all files Collapse all files

24
agents/imitation_agent.py
View File

@@ -13,23 +13,27 @@
# See the License for the specific language governing permissions and
# limitations under the License.
#
import collections
from agents.agent import *
from agents import agent
from architectures import network_wrapper as nw
import utils
import logging
# Imitation Agent
class ImitationAgent(Agent):
class ImitationAgent(agent.Agent):
def __init__(self, env, tuning_parameters, replicated_device=None, thread_id=0):
Agent.__init__(self, env, tuning_parameters, replicated_device, thread_id)
self.main_network = NetworkWrapper(tuning_parameters, False, self.has_global, 'main',
self.replicated_device, self.worker_device)
agent.Agent.__init__(self, env, tuning_parameters, replicated_device, thread_id)
self.main_network = nw.NetworkWrapper(tuning_parameters, False, self.has_global, 'main',
self.replicated_device, self.worker_device)
self.networks.append(self.main_network)
self.imitation = True
def extract_action_values(self, prediction):
return prediction.squeeze()
def choose_action(self, curr_state, phase=RunPhase.TRAIN):
def choose_action(self, curr_state, phase=utils.RunPhase.TRAIN):
# convert to batch so we can run it through the network
prediction = self.main_network.online_network.predict(self.tf_input_state(curr_state))
@@ -49,10 +53,10 @@ class ImitationAgent(Agent):
def log_to_screen(self, phase):
# log to screen
if phase == RunPhase.TRAIN:
if phase == utils.RunPhase.TRAIN:
# for the training phase - we log during the episode to visualize the progress in training
screen.log_dict(
OrderedDict([
logging.screen.log_dict(
collections.OrderedDict([
("Worker", self.task_id),
("Episode", self.current_episode),
("Loss", self.loss.values[-1]),
@@ -62,4 +66,4 @@ class ImitationAgent(Agent):
)
else:
# for the evaluation phase - logging as in regular RL
Agent.log_to_screen(self, phase)
agent.Agent.log_to_screen(self, phase)