mirror of
https://github.com/saymrwulf/stable-baselines3.git
synced 2026-11-18 09:54:10 +00:00
Reformated every file with PEP 8 errors
This commit is contained in:
parent
b20b70db48
commit
a7655ca6e1
10 changed files with 62 additions and 65 deletions
12
docs/conf.py
12
docs/conf.py
|
|
@ -16,12 +16,14 @@ import os
|
||||||
import sys
|
import sys
|
||||||
from unittest.mock import MagicMock
|
from unittest.mock import MagicMock
|
||||||
|
|
||||||
|
import torchy_baselines
|
||||||
# source code directory, relative to this file, for sphinx-autobuild
|
# source code directory, relative to this file, for sphinx-autobuild
|
||||||
sys.path.insert(0, os.path.abspath('..'))
|
sys.path.insert(0, os.path.abspath('..'))
|
||||||
|
|
||||||
|
|
||||||
class Mock(MagicMock):
|
class Mock(MagicMock):
|
||||||
__subclasses__ = []
|
__subclasses__ = []
|
||||||
|
|
||||||
@classmethod
|
@classmethod
|
||||||
def __getattr__(cls, name):
|
def __getattr__(cls, name):
|
||||||
return MagicMock()
|
return MagicMock()
|
||||||
|
|
@ -42,10 +44,6 @@ MOCK_MODULES = ['joblib', 'scipy', 'scipy.signal',
|
||||||
'gym.wrappers', 'gym.wrappers.monitoring', 'zmq']
|
'gym.wrappers', 'gym.wrappers.monitoring', 'zmq']
|
||||||
sys.modules.update((mod_name, Mock()) for mod_name in MOCK_MODULES)
|
sys.modules.update((mod_name, Mock()) for mod_name in MOCK_MODULES)
|
||||||
|
|
||||||
|
|
||||||
import torchy_baselines
|
|
||||||
|
|
||||||
|
|
||||||
# -- Project information -----------------------------------------------------
|
# -- Project information -----------------------------------------------------
|
||||||
|
|
||||||
project = 'Torchy Baselines'
|
project = 'Torchy Baselines'
|
||||||
|
|
@ -57,7 +55,6 @@ version = 'master (' + torchy_baselines.__version__ + ' )'
|
||||||
# The full version, including alpha/beta/rc tags
|
# The full version, including alpha/beta/rc tags
|
||||||
release = torchy_baselines.__version__
|
release = torchy_baselines.__version__
|
||||||
|
|
||||||
|
|
||||||
# -- General configuration ---------------------------------------------------
|
# -- General configuration ---------------------------------------------------
|
||||||
|
|
||||||
# If your documentation needs a minimal Sphinx version, state it here.
|
# If your documentation needs a minimal Sphinx version, state it here.
|
||||||
|
|
@ -102,7 +99,6 @@ exclude_patterns = ['_build', 'Thumbs.db', '.DS_Store']
|
||||||
# The name of the Pygments (syntax highlighting) style to use.
|
# The name of the Pygments (syntax highlighting) style to use.
|
||||||
pygments_style = 'sphinx'
|
pygments_style = 'sphinx'
|
||||||
|
|
||||||
|
|
||||||
# -- Options for HTML output -------------------------------------------------
|
# -- Options for HTML output -------------------------------------------------
|
||||||
|
|
||||||
# The theme to use for HTML and HTML Help pages. See the documentation for
|
# The theme to use for HTML and HTML Help pages. See the documentation for
|
||||||
|
|
@ -121,6 +117,7 @@ html_logo = '_static/img/logo.png'
|
||||||
def setup(app):
|
def setup(app):
|
||||||
app.add_stylesheet("css/baselines_theme.css")
|
app.add_stylesheet("css/baselines_theme.css")
|
||||||
|
|
||||||
|
|
||||||
# Theme options are theme-specific and customize the look and feel of a theme
|
# Theme options are theme-specific and customize the look and feel of a theme
|
||||||
# further. For a list of options available for each theme, see the
|
# further. For a list of options available for each theme, see the
|
||||||
# documentation.
|
# documentation.
|
||||||
|
|
@ -148,7 +145,6 @@ html_static_path = ['_static']
|
||||||
# Output file base name for HTML help builder.
|
# Output file base name for HTML help builder.
|
||||||
htmlhelp_basename = 'TorchyBaselinesdoc'
|
htmlhelp_basename = 'TorchyBaselinesdoc'
|
||||||
|
|
||||||
|
|
||||||
# -- Options for LaTeX output ------------------------------------------------
|
# -- Options for LaTeX output ------------------------------------------------
|
||||||
|
|
||||||
latex_elements = {
|
latex_elements = {
|
||||||
|
|
@ -177,7 +173,6 @@ latex_documents = [
|
||||||
'Torchy Baselines Contributors', 'manual'),
|
'Torchy Baselines Contributors', 'manual'),
|
||||||
]
|
]
|
||||||
|
|
||||||
|
|
||||||
# -- Options for manual page output ------------------------------------------
|
# -- Options for manual page output ------------------------------------------
|
||||||
|
|
||||||
# One entry per manual page. List of tuples
|
# One entry per manual page. List of tuples
|
||||||
|
|
@ -187,7 +182,6 @@ man_pages = [
|
||||||
[author], 1)
|
[author], 1)
|
||||||
]
|
]
|
||||||
|
|
||||||
|
|
||||||
# -- Options for Texinfo output ----------------------------------------------
|
# -- Options for Texinfo output ----------------------------------------------
|
||||||
|
|
||||||
# Grouping the document tree into Texinfo files. List of tuples
|
# Grouping the document tree into Texinfo files. List of tuples
|
||||||
|
|
|
||||||
|
|
@ -5,6 +5,7 @@ import pytest
|
||||||
|
|
||||||
from torchy_baselines import PPO
|
from torchy_baselines import PPO
|
||||||
|
|
||||||
|
|
||||||
@pytest.mark.parametrize('net_arch', [
|
@pytest.mark.parametrize('net_arch', [
|
||||||
[12, dict(vf=[16], pi=[8])],
|
[12, dict(vf=[16], pi=[8])],
|
||||||
[4],
|
[4],
|
||||||
|
|
|
||||||
|
|
@ -6,7 +6,6 @@ import numpy as np
|
||||||
from torchy_baselines import A2C, CEMRL, PPO, SAC, TD3
|
from torchy_baselines import A2C, CEMRL, PPO, SAC, TD3
|
||||||
from torchy_baselines.common.noise import NormalActionNoise
|
from torchy_baselines.common.noise import NormalActionNoise
|
||||||
|
|
||||||
|
|
||||||
action_noise = NormalActionNoise(np.zeros(1), 0.1 * np.ones(1))
|
action_noise = NormalActionNoise(np.zeros(1), 0.1 * np.ones(1))
|
||||||
|
|
||||||
|
|
||||||
|
|
@ -19,8 +18,6 @@ def test_td3():
|
||||||
os.remove("test_save.pth")
|
os.remove("test_save.pth")
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
def test_a2c():
|
def test_a2c():
|
||||||
model = A2C('MlpPolicy', 'CartPole-v1', policy_kwargs=dict(net_arch=[16]), verbose=1, create_eval_env=True)
|
model = A2C('MlpPolicy', 'CartPole-v1', policy_kwargs=dict(net_arch=[16]), verbose=1, create_eval_env=True)
|
||||||
model.learn(total_timesteps=1000, eval_freq=500)
|
model.learn(total_timesteps=1000, eval_freq=500)
|
||||||
|
|
@ -47,6 +44,7 @@ def test_onpolicy(model_class, env_id):
|
||||||
# model.load("test_save")
|
# model.load("test_save")
|
||||||
# os.remove("test_save.pth")
|
# os.remove("test_save.pth")
|
||||||
|
|
||||||
|
|
||||||
def test_sac():
|
def test_sac():
|
||||||
model = SAC('MlpPolicy', 'Pendulum-v0', policy_kwargs=dict(net_arch=[64, 64]),
|
model = SAC('MlpPolicy', 'Pendulum-v0', policy_kwargs=dict(net_arch=[64, 64]),
|
||||||
learning_starts=100, verbose=1, create_eval_env=True, ent_coef='auto',
|
learning_starts=100, verbose=1, create_eval_env=True, ent_coef='auto',
|
||||||
|
|
|
||||||
|
|
@ -59,8 +59,10 @@ class CustomGymEnv(gym.Env):
|
||||||
@pytest.mark.parametrize('vec_env_wrapper', VEC_ENV_WRAPPERS)
|
@pytest.mark.parametrize('vec_env_wrapper', VEC_ENV_WRAPPERS)
|
||||||
def test_vecenv_custom_calls(vec_env_class, vec_env_wrapper):
|
def test_vecenv_custom_calls(vec_env_class, vec_env_wrapper):
|
||||||
"""Test access to methods/attributes of vectorized environments"""
|
"""Test access to methods/attributes of vectorized environments"""
|
||||||
|
|
||||||
def make_env():
|
def make_env():
|
||||||
return CustomGymEnv(gym.spaces.Box(low=np.zeros(2), high=np.ones(2)))
|
return CustomGymEnv(gym.spaces.Box(low=np.zeros(2), high=np.ones(2)))
|
||||||
|
|
||||||
vec_env = vec_env_class([make_env for _ in range(N_ENVS)])
|
vec_env = vec_env_class([make_env for _ in range(N_ENVS)])
|
||||||
|
|
||||||
if vec_env_wrapper is not None:
|
if vec_env_wrapper is not None:
|
||||||
|
|
@ -92,7 +94,6 @@ def test_vecenv_custom_calls(vec_env_class, vec_env_wrapper):
|
||||||
assert (env_method_subset[1] == np.ones((1, 3))).all()
|
assert (env_method_subset[1] == np.ones((1, 3))).all()
|
||||||
assert len(env_method_subset) == 2
|
assert len(env_method_subset) == 2
|
||||||
|
|
||||||
|
|
||||||
# Test to change value for all the environments
|
# Test to change value for all the environments
|
||||||
setattr_result = vec_env.set_attr('current_step', 42, indices=None)
|
setattr_result = vec_env.set_attr('current_step', 42, indices=None)
|
||||||
getattr_result = vec_env.get_attr('current_step')
|
getattr_result = vec_env.get_attr('current_step')
|
||||||
|
|
@ -193,8 +194,10 @@ SPACES = collections.OrderedDict([
|
||||||
('continuous', gym.spaces.Box(low=np.zeros(2), high=np.ones(2))),
|
('continuous', gym.spaces.Box(low=np.zeros(2), high=np.ones(2))),
|
||||||
])
|
])
|
||||||
|
|
||||||
|
|
||||||
def check_vecenv_spaces(vec_env_class, space, obs_assert):
|
def check_vecenv_spaces(vec_env_class, space, obs_assert):
|
||||||
"""Helper method to check observation spaces in vectorized environments."""
|
"""Helper method to check observation spaces in vectorized environments."""
|
||||||
|
|
||||||
def make_env():
|
def make_env():
|
||||||
return CustomGymEnv(space)
|
return CustomGymEnv(space)
|
||||||
|
|
||||||
|
|
@ -228,6 +231,7 @@ def test_vecenv_single_space(vec_env_class, space):
|
||||||
|
|
||||||
class _UnorderedDictSpace(gym.spaces.Dict):
|
class _UnorderedDictSpace(gym.spaces.Dict):
|
||||||
"""Like DictSpace, but returns an unordered dict when sampling."""
|
"""Like DictSpace, but returns an unordered dict when sampling."""
|
||||||
|
|
||||||
def sample(self):
|
def sample(self):
|
||||||
return dict(super().sample())
|
return dict(super().sample())
|
||||||
|
|
||||||
|
|
@ -301,14 +305,17 @@ class CustomWrapperB(VecNormalize):
|
||||||
def name_test(self):
|
def name_test(self):
|
||||||
return self.__class__
|
return self.__class__
|
||||||
|
|
||||||
|
|
||||||
class CustomWrapperBB(CustomWrapperB):
|
class CustomWrapperBB(CustomWrapperB):
|
||||||
def __init__(self, venv):
|
def __init__(self, venv):
|
||||||
CustomWrapperB.__init__(self, venv)
|
CustomWrapperB.__init__(self, venv)
|
||||||
self.var_bb = 'bb'
|
self.var_bb = 'bb'
|
||||||
|
|
||||||
|
|
||||||
def test_vecenv_wrapper_getattr():
|
def test_vecenv_wrapper_getattr():
|
||||||
def make_env():
|
def make_env():
|
||||||
return CustomGymEnv(gym.spaces.Box(low=np.zeros(2), high=np.ones(2)))
|
return CustomGymEnv(gym.spaces.Box(low=np.zeros(2), high=np.ones(2)))
|
||||||
|
|
||||||
vec_env = DummyVecEnv([make_env for _ in range(N_ENVS)])
|
vec_env = DummyVecEnv([make_env for _ in range(N_ENVS)])
|
||||||
wrapped = CustomWrapperA(CustomWrapperBB(vec_env))
|
wrapped = CustomWrapperA(CustomWrapperBB(vec_env))
|
||||||
assert wrapped.var_a == 'a'
|
assert wrapped.var_a == 'a'
|
||||||
|
|
|
||||||
|
|
@ -115,15 +115,15 @@ class A2C(PPO):
|
||||||
self.policy.optimizer.step()
|
self.policy.optimizer.step()
|
||||||
# approx_kl_divs.append(th.mean(old_log_prob - log_prob).detach().cpu().numpy())
|
# approx_kl_divs.append(th.mean(old_log_prob - log_prob).detach().cpu().numpy())
|
||||||
|
|
||||||
# print(explained_variance(self.rollout_buffer.returns.flatten().cpu().numpy(),
|
# print(explained_variance(self.rollout_buffer.returns.flatten().cpu().numpy(),
|
||||||
# self.rollout_buffer.values.flatten().cpu().numpy()))
|
# self.rollout_buffer.values.flatten().cpu().numpy()))
|
||||||
|
|
||||||
def learn(self, total_timesteps, callback=None, log_interval=100,
|
def learn(self, total_timesteps, callback=None, log_interval=100,
|
||||||
eval_env=None, eval_freq=-1, n_eval_episodes=5, tb_log_name="A2C", reset_num_timesteps=True):
|
eval_env=None, eval_freq=-1, n_eval_episodes=5, tb_log_name="A2C", reset_num_timesteps=True):
|
||||||
|
|
||||||
return super(A2C, self).learn(total_timesteps=total_timesteps, callback=callback, log_interval=log_interval,
|
return super(A2C, self).learn(total_timesteps=total_timesteps, callback=callback, log_interval=log_interval,
|
||||||
eval_env=eval_env, eval_freq=eval_freq, n_eval_episodes=n_eval_episodes,
|
eval_env=eval_env, eval_freq=eval_freq, n_eval_episodes=n_eval_episodes,
|
||||||
tb_log_name=tb_log_name, reset_num_timesteps=reset_num_timesteps)
|
tb_log_name=tb_log_name, reset_num_timesteps=reset_num_timesteps)
|
||||||
|
|
||||||
def save(self, path):
|
def save(self, path):
|
||||||
if not path.endswith('.pth'):
|
if not path.endswith('.pth'):
|
||||||
|
|
|
||||||
|
|
@ -290,9 +290,10 @@ class BaseRLModel(object):
|
||||||
|
|
||||||
if 'policy_kwargs' in kwargs and kwargs['policy_kwargs'] != data['policy_kwargs']:
|
if 'policy_kwargs' in kwargs and kwargs['policy_kwargs'] != data['policy_kwargs']:
|
||||||
raise ValueError("The specified policy kwargs do not equal the stored policy kwargs."
|
raise ValueError("The specified policy kwargs do not equal the stored policy kwargs."
|
||||||
"Stored kwargs: {}, specified kwargs: {}".format(data['policy_kwargs'],kwargs['policy_kwargs']))
|
"Stored kwargs: {}, specified kwargs: {}".format(data['policy_kwargs'],
|
||||||
|
kwargs['policy_kwargs']))
|
||||||
|
|
||||||
model = cls(policy=data["policy"],env=None, _init_setup_model=False)
|
model = cls(policy=data["policy"], env=None, _init_setup_model=False)
|
||||||
model.__dict__.update(data)
|
model.__dict__.update(data)
|
||||||
model.__dict__.update(kwargs)
|
model.__dict__.update(kwargs)
|
||||||
model.set_env(env)
|
model.set_env(env)
|
||||||
|
|
@ -300,40 +301,8 @@ class BaseRLModel(object):
|
||||||
|
|
||||||
return model
|
return model
|
||||||
|
|
||||||
|
|
||||||
@staticmethod
|
@staticmethod
|
||||||
def _save_to_file_zip(save_path, data=None, params=None):
|
def _load_from_file(load_path, load_data=True):
|
||||||
"""Save model to a zip archive
|
|
||||||
|
|
||||||
:param save_path: (str or file-like) Where to store the model
|
|
||||||
:param data: (OrderedDict) Class parameters being stored
|
|
||||||
:param params: (OrderedDict) Model parameters being stored expexted to be state_dict
|
|
||||||
"""
|
|
||||||
|
|
||||||
# data/params can be None, so do not
|
|
||||||
# try to serialize them blindly
|
|
||||||
if data is not None:
|
|
||||||
serialized_data = data_to_json(data)
|
|
||||||
|
|
||||||
# Check postfix if save_path is a string
|
|
||||||
if isinstance(save_path, str):
|
|
||||||
_, ext = os.path.splitext(save_path)
|
|
||||||
if ext == "":
|
|
||||||
save_path += ".zip"
|
|
||||||
|
|
||||||
# Create a zip-archive and write our objects
|
|
||||||
# there. This works when save_path is either
|
|
||||||
# str or a file-like
|
|
||||||
with zipfile.ZipFile(save_path, "w") as file_:
|
|
||||||
# Do not try to save "None" elements
|
|
||||||
if data is not None:
|
|
||||||
file_.writestr("data",serialized_data)
|
|
||||||
if params is not None:
|
|
||||||
with file_.open('param.pth', mode="w") as param_file:
|
|
||||||
th.save(params,param_file)
|
|
||||||
|
|
||||||
@staticmethod
|
|
||||||
def _load_from_file(load_path, load_data = True):
|
|
||||||
""" Load model data from a .zip archive
|
""" Load model data from a .zip archive
|
||||||
|
|
||||||
:param load_path: (str or file-like) Where to load the model from
|
:param load_path: (str or file-like) Where to load the model from
|
||||||
|
|
@ -351,7 +320,7 @@ class BaseRLModel(object):
|
||||||
|
|
||||||
# Open the zip archive and load data
|
# Open the zip archive and load data
|
||||||
try:
|
try:
|
||||||
with zipfile.ZipFile(load_path,"r") as file_:
|
with zipfile.ZipFile(load_path, "r") as file_:
|
||||||
namelist = file_.namelist()
|
namelist = file_.namelist()
|
||||||
# If data or parameters is not in the
|
# If data or parameters is not in the
|
||||||
# zip archive, assume they were stored
|
# zip archive, assume they were stored
|
||||||
|
|
@ -378,12 +347,6 @@ class BaseRLModel(object):
|
||||||
|
|
||||||
return data, params
|
return data, params
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
def set_random_seed(self, seed=0):
|
def set_random_seed(self, seed=0):
|
||||||
set_random_seed(seed, using_cuda=self.device == th.device('cuda'))
|
set_random_seed(seed, using_cuda=self.device == th.device('cuda'))
|
||||||
self.action_space.seed(seed)
|
self.action_space.seed(seed)
|
||||||
|
|
@ -478,7 +441,7 @@ class BaseRLModel(object):
|
||||||
|
|
||||||
# Display training infos
|
# Display training infos
|
||||||
if self.verbose >= 1 and log_interval is not None and (
|
if self.verbose >= 1 and log_interval is not None and (
|
||||||
episode_num + total_episodes) % log_interval == 0:
|
episode_num + total_episodes) % log_interval == 0:
|
||||||
fps = int(num_timesteps / (time.time() - self.start_time))
|
fps = int(num_timesteps / (time.time() - self.start_time))
|
||||||
logger.logkv("episodes", episode_num + total_episodes)
|
logger.logkv("episodes", episode_num + total_episodes)
|
||||||
# logger.logkv("mean 100 episode reward", mean_reward)
|
# logger.logkv("mean 100 episode reward", mean_reward)
|
||||||
|
|
@ -495,3 +458,34 @@ class BaseRLModel(object):
|
||||||
mean_reward = np.mean(episode_rewards) if total_episodes > 0 else 0.0
|
mean_reward = np.mean(episode_rewards) if total_episodes > 0 else 0.0
|
||||||
|
|
||||||
return mean_reward, total_steps, total_episodes, obs
|
return mean_reward, total_steps, total_episodes, obs
|
||||||
|
|
||||||
|
|
||||||
|
def _save_to_file_zip(save_path, data=None, params=None):
|
||||||
|
"""Save model to a zip archive
|
||||||
|
|
||||||
|
:param save_path: (str or file-like) Where to store the model
|
||||||
|
:param data: (OrderedDict) Class parameters being stored
|
||||||
|
:param params: (OrderedDict) Model parameters being stored expexted to be state_dict
|
||||||
|
"""
|
||||||
|
|
||||||
|
# data/params can be None, so do not
|
||||||
|
# try to serialize them blindly
|
||||||
|
if data is not None:
|
||||||
|
serialized_data = data_to_json(data)
|
||||||
|
|
||||||
|
# Check postfix if save_path is a string
|
||||||
|
if isinstance(save_path, str):
|
||||||
|
_, ext = os.path.splitext(save_path)
|
||||||
|
if ext == "":
|
||||||
|
save_path += ".zip"
|
||||||
|
|
||||||
|
# Create a zip-archive and write our objects
|
||||||
|
# there. This works when save_path is either
|
||||||
|
# str or a file-like
|
||||||
|
with zipfile.ZipFile(save_path, "w") as file_:
|
||||||
|
# Do not try to save "None" elements
|
||||||
|
if data is not None:
|
||||||
|
file_.writestr("data", serialized_data)
|
||||||
|
if params is not None:
|
||||||
|
with file_.open('param.pth', mode="w") as param_file:
|
||||||
|
th.save(params, param_file)
|
||||||
|
|
|
||||||
|
|
@ -4,6 +4,7 @@ import torch.nn as nn
|
||||||
from torch.distributions import Normal, Categorical
|
from torch.distributions import Normal, Categorical
|
||||||
from gym import spaces
|
from gym import spaces
|
||||||
|
|
||||||
|
|
||||||
class Distribution(object):
|
class Distribution(object):
|
||||||
def __init__(self):
|
def __init__(self):
|
||||||
super(Distribution, self).__init__()
|
super(Distribution, self).__init__()
|
||||||
|
|
@ -97,7 +98,8 @@ class SquashedDiagGaussianDistribution(DiagGaussianDistribution):
|
||||||
self.gaussian_action = None
|
self.gaussian_action = None
|
||||||
|
|
||||||
def proba_distribution(self, mean_actions, log_std, deterministic=False):
|
def proba_distribution(self, mean_actions, log_std, deterministic=False):
|
||||||
action, _ = super(SquashedDiagGaussianDistribution, self).proba_distribution(mean_actions, log_std, deterministic)
|
action, _ = super(SquashedDiagGaussianDistribution, self).proba_distribution(mean_actions, log_std,
|
||||||
|
deterministic)
|
||||||
return action, self
|
return action, self
|
||||||
|
|
||||||
def mode(self):
|
def mode(self):
|
||||||
|
|
|
||||||
|
|
@ -42,6 +42,7 @@ def _worker(remote, parent_remote, env_fn_wrapper):
|
||||||
except EOFError:
|
except EOFError:
|
||||||
break
|
break
|
||||||
|
|
||||||
|
|
||||||
def tile_images(img_nhwc):
|
def tile_images(img_nhwc):
|
||||||
"""
|
"""
|
||||||
Tile N images into one big PxQ image
|
Tile N images into one big PxQ image
|
||||||
|
|
@ -68,7 +69,6 @@ def tile_images(img_nhwc):
|
||||||
return out_image
|
return out_image
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
class SubprocVecEnv(VecEnv):
|
class SubprocVecEnv(VecEnv):
|
||||||
"""
|
"""
|
||||||
Creates a multiprocess vectorized wrapper for multiple environments, distributing each environment to its own
|
Creates a multiprocess vectorized wrapper for multiple environments, distributing each environment to its own
|
||||||
|
|
|
||||||
|
|
@ -321,7 +321,7 @@ class PPO(BaseRLModel):
|
||||||
|
|
||||||
params_to_save = self.get_policy_parameters()
|
params_to_save = self.get_policy_parameters()
|
||||||
|
|
||||||
self._save_to_file_zip(path, data=data, params=params_to_save)
|
_save_to_file_zip(path, data=data, params=params_to_save)
|
||||||
|
|
||||||
"""def load(self, path, env=None, **_kwargs):
|
"""def load(self, path, env=None, **_kwargs):
|
||||||
if not path.endswith('.pth'):
|
if not path.endswith('.pth'):
|
||||||
|
|
|
||||||
|
|
@ -148,7 +148,8 @@ class TD3(BaseRLModel):
|
||||||
for param, target_param in zip(self.critic.parameters(), self.critic_target.parameters()):
|
for param, target_param in zip(self.critic.parameters(), self.critic_target.parameters()):
|
||||||
target_param.data.copy_(tau * param.data + (1 - tau) * target_param.data)
|
target_param.data.copy_(tau * param.data + (1 - tau) * target_param.data)
|
||||||
|
|
||||||
def train_actor(self, gradient_steps=1, batch_size=100, tau_actor=0.005, tau_critic=0.005, replay_data=None):
|
def train_actor(self, gradient_steps: object = 1, batch_size: object = 100, tau_actor: object = 0.005, tau_critic: object = 0.005,
|
||||||
|
replay_data: object = None) -> object:
|
||||||
# Update optimizer learning rate
|
# Update optimizer learning rate
|
||||||
self._update_learning_rate(self.actor.optimizer)
|
self._update_learning_rate(self.actor.optimizer)
|
||||||
|
|
||||||
|
|
|
||||||
Loading…
Reference in a new issue