mirror of
https://github.com/wassname/DeepRL.git
synced 2026-09-10 11:40:58 +08:00
Refactor tasks
This commit is contained in:
Shangtong Zhang
+1
-1
@@ -11,7 +11,7 @@ data
|
||||
dataset
|
||||
draw_*
|
||||
log
|
||||
evaluation_log
|
||||
old_logs
|
||||
figure
|
||||
to_plot
|
||||
|
||||
|
||||
+47
-84
@@ -14,77 +14,55 @@ from utils import *
|
||||
import datetime
|
||||
import uuid
|
||||
|
||||
class BasicTask:
|
||||
def __init__(self, max_steps=sys.maxsize):
|
||||
self.steps = 0
|
||||
self.max_steps = max_steps
|
||||
class BaseTask:
|
||||
def set_monitor(self, env, log_dir):
|
||||
if log_dir is None:
|
||||
return env
|
||||
mkdir(log_dir)
|
||||
return Monitor(env, '%s/%s' % (log_dir, uuid.uuid1()))
|
||||
|
||||
def reset(self):
|
||||
self.steps = 0
|
||||
state = self.env.reset()
|
||||
return state
|
||||
return self.env.reset()
|
||||
|
||||
def step(self, action):
|
||||
next_state, reward, done, info = self.env.step(action)
|
||||
self.steps += 1
|
||||
done = (done or self.steps >= self.max_steps)
|
||||
return next_state, reward, done, info
|
||||
return self.env.step(action)
|
||||
|
||||
class ClassicalControl(BasicTask):
|
||||
def seed(self, random_seed):
|
||||
return self.env.seed(random_seed)
|
||||
|
||||
class ClassicalControl(BaseTask):
|
||||
def __init__(self, name='CartPole-v0', max_steps=200, log_dir=None):
|
||||
BasicTask.__init__(self, max_steps)
|
||||
BaseTask.__init__(self)
|
||||
self.name = name
|
||||
self.env = gym.make(self.name)
|
||||
self.env._max_episode_steps = sys.maxsize
|
||||
self.env._max_episode_steps = max_steps
|
||||
self.action_dim = self.env.action_space.n
|
||||
self.state_dim = self.env.observation_space.shape[0]
|
||||
if log_dir is not None:
|
||||
mkdir(log_dir)
|
||||
self.env = Monitor(self.env, '%s/%s' % (log_dir, uuid.uuid1()))
|
||||
self.env = self.set_monitor(self.env, log_dir)
|
||||
|
||||
class LunarLander(BasicTask):
|
||||
name = 'LunarLander-v2'
|
||||
success_threshold = 200
|
||||
|
||||
def __init__(self, max_steps=sys.maxsize, log_dir=None):
|
||||
BasicTask.__init__(self, max_steps)
|
||||
self.env = gym.make(self.name)
|
||||
self.action_dim = self.env.action_space.n
|
||||
self.state_dim = self.env.observation_space.shape[0]
|
||||
if log_dir is not None:
|
||||
mkdir(log_dir)
|
||||
self.env = Monitor(self.env, '%s/%s' % (log_dir, uuid.uuid1()))
|
||||
|
||||
class PixelAtari(BasicTask):
|
||||
def __init__(self, name, seed=0, log_dir=None, max_steps=sys.maxsize,
|
||||
class PixelAtari(BaseTask):
|
||||
def __init__(self, name, seed=0, log_dir=None,
|
||||
frame_skip=4, history_length=4, dataset=False):
|
||||
BasicTask.__init__(self, max_steps)
|
||||
BaseTask.__init__(self)
|
||||
env = make_atari(name, frame_skip)
|
||||
env.seed(seed)
|
||||
if dataset:
|
||||
env = DatasetEnv(env)
|
||||
self.dataset_env = env
|
||||
if log_dir is not None:
|
||||
mkdir(log_dir)
|
||||
env = Monitor(env, '%s/%s' % (log_dir, uuid.uuid1()))
|
||||
env = self.set_monitor(env, log_dir)
|
||||
env = wrap_deepmind(env, history_length=history_length)
|
||||
self.env = env
|
||||
self.action_dim = self.env.action_space.n
|
||||
self.state_dim = self.env.observation_space.shape
|
||||
self.name = name
|
||||
|
||||
def normalize_state(self, state):
|
||||
return np.asarray(state) / 255.0
|
||||
|
||||
class RamAtari(BasicTask):
|
||||
def __init__(self, name, no_op, frame_skip, max_steps=sys.maxsize, log_dir=None):
|
||||
BasicTask.__init__(self, max_steps)
|
||||
class RamAtari(BaseTask):
|
||||
def __init__(self, name, no_op, frame_skip, log_dir=None):
|
||||
BaseTask.__init__(self)
|
||||
self.name = name
|
||||
env = gym.make(name)
|
||||
assert 'NoFrameskip' in env.spec.id
|
||||
if log_dir is not None:
|
||||
mkdir(log_dir)
|
||||
env = Monitor(env, '%s/%s' % (log_dir, uuid.uuid1()))
|
||||
env = self.set_monitor(env, log_dir)
|
||||
env = EpisodicLifeEnv(env)
|
||||
env = NoopResetEnv(env, noop_max=no_op)
|
||||
env = SkipEnv(env, skip=frame_skip)
|
||||
@@ -94,81 +72,66 @@ class RamAtari(BasicTask):
|
||||
self.action_dim = self.env.action_space.n
|
||||
self.state_dim = 128
|
||||
|
||||
def normalize_state(self, state):
|
||||
return np.asarray(state) / 255.0
|
||||
|
||||
class Pendulum(BasicTask):
|
||||
name = 'Pendulum-v0'
|
||||
success_threshold = -10
|
||||
|
||||
def __init__(self, max_steps=sys.maxsize, log_dir=None):
|
||||
BasicTask.__init__(self, max_steps)
|
||||
class Pendulum(BaseTask):
|
||||
def __init__(self, log_dir=None):
|
||||
BaseTask.__init__(self)
|
||||
self.name = 'Pendulum-v0'
|
||||
self.env = gym.make(self.name)
|
||||
self.action_dim = self.env.action_space.shape[0]
|
||||
self.state_dim = self.env.observation_space.shape[0]
|
||||
if log_dir is not None:
|
||||
mkdir(log_dir)
|
||||
self.env = Monitor(self.env, '%s/%s' % (log_dir, uuid.uuid1()))
|
||||
self.env = self.set_monitor(self.env, log_dir)
|
||||
|
||||
def step(self, action):
|
||||
return BasicTask.step(self, np.clip(2 * action, -2, 2))
|
||||
return BaseTask.step(self, np.clip(2 * action, -2, 2))
|
||||
|
||||
class Box2DContinuous(BasicTask):
|
||||
def __init__(self, name, max_steps=sys.maxsize, log_dir=None):
|
||||
BasicTask.__init__(self, max_steps)
|
||||
class Box2DContinuous(BaseTask):
|
||||
def __init__(self, name, log_dir=None):
|
||||
BaseTask.__init__(self)
|
||||
self.name = name
|
||||
self.env = gym.make(self.name)
|
||||
self.action_dim = self.env.action_space.shape[0]
|
||||
self.state_dim = self.env.observation_space.shape[0]
|
||||
if log_dir is not None:
|
||||
mkdir(log_dir)
|
||||
self.env = Monitor(self.env, '%s/%s' % (log_dir, uuid.uuid1()))
|
||||
self.env = self.set_monitor(self.env, log_dir)
|
||||
|
||||
def step(self, action):
|
||||
return BasicTask.step(self, np.clip(action, -1, 1))
|
||||
return BaseTask.step(self, np.clip(action, -1, 1))
|
||||
|
||||
class Roboschool(BasicTask):
|
||||
def __init__(self, name, max_steps=sys.maxsize, log_dir=None):
|
||||
class Roboschool(BaseTask):
|
||||
def __init__(self, name, log_dir=None):
|
||||
import roboschool
|
||||
BasicTask.__init__(self, max_steps)
|
||||
BaseTask.__init__(self)
|
||||
self.name = name
|
||||
self.env = gym.make(self.name)
|
||||
self.action_dim = self.env.action_space.shape[0]
|
||||
self.state_dim = self.env.observation_space.shape[0]
|
||||
if log_dir is not None:
|
||||
mkdir(log_dir)
|
||||
self.env = Monitor(self.env, '%s/%s' % (log_dir, uuid.uuid1()))
|
||||
self.env = self.set_monitor(self.env, log_dir)
|
||||
|
||||
def step(self, action):
|
||||
return BasicTask.step(self, np.clip(action, -1, 1))
|
||||
return BaseTask.step(self, np.clip(action, -1, 1))
|
||||
|
||||
class DMControl(BasicTask):
|
||||
def __init__(self, domain_name, task_name, max_steps=sys.maxsize, log_dir=None):
|
||||
class DMControl(BaseTask):
|
||||
def __init__(self, domain_name, task_name, log_dir=None):
|
||||
from dm_control import suite
|
||||
import dm_control2gym
|
||||
BasicTask.__init__(self, max_steps)
|
||||
BaseTask.__init__(self)
|
||||
|
||||
self.name = domain_name + '_' + task_name
|
||||
self.env = dm_control2gym.make(domain_name, task_name)
|
||||
|
||||
self.action_dim = self.env.action_space.shape[0]
|
||||
self.state_dim = self.env.observation_space.shape[0]
|
||||
if log_dir is not None:
|
||||
mkdir(log_dir)
|
||||
self.env = Monitor(self.env, '%s/%s' % (log_dir, uuid.uuid1()))
|
||||
self.env = self.set_monitor(self.env, log_dir)
|
||||
|
||||
class GymRobotics(BasicTask):
|
||||
class GymRobotics(BaseTask):
|
||||
def __init__(self, name, log_dir=None):
|
||||
BasicTask.__init__(self)
|
||||
BaseTask.__init__(self)
|
||||
|
||||
self.name = name
|
||||
self.env = gym.make(name)
|
||||
|
||||
self.action_dim = self.env.action_space.shape[0]
|
||||
self.state_dim = len(self.flatten_state(self.env.reset()))
|
||||
if log_dir is not None:
|
||||
mkdir(log_dir)
|
||||
self.env = Monitor(self.env, '%s/%s' % (log_dir, uuid.uuid1()))
|
||||
self.env = self.set_monitor(self.env, log_dir)
|
||||
|
||||
def flatten_state(self, state):
|
||||
flat = []
|
||||
@@ -189,7 +152,7 @@ def sub_task(parent_pipe, pipe, task_fn, rank, log_dir):
|
||||
seed = np.random.randint(0, sys.maxsize)
|
||||
parent_pipe.close()
|
||||
task = task_fn(log_dir=log_dir)
|
||||
task.env.seed(seed)
|
||||
task.seed(seed)
|
||||
while True:
|
||||
op, data = pipe.recv()
|
||||
if op == 'step':
|
||||
|
||||
Reference in New Issue
Block a user