Fix some typos. (#1095)

Signed-off-by: Yulv-git <[email protected]>

benchmark/fluid/MADDPG/train.py

@@ -123,7 +123,7 @@ def train_agent():
     final_ep_ag_rewards = []  # agent rewards for training curve
 
     if args.restore:
-        # restore modle
+        # restore model
         for i in range(len(agents)):
             model_file = args.model_dir + '/agent_' + str(i)
             if not os.path.exists(model_file):

benchmark/fluid/offline-Q-learning/README.md

@@ -18,7 +18,7 @@ learn_program = parl.compile(learn_program, loss=training_loss)
 
 ## Demonstration
 
-We provide a demonstration of offline Q-learning with parallel executing, in which we seperate the procedures of collecting data and training the model. First we collect data by interacting with the environment and save them to a replay memory file, and then fit and evaluate the Q network with the collected data. Repeat these two steps to improve the performance gradually.
+We provide a demonstration of offline Q-learning with parallel executing, in which we separate the procedures of collecting data and training the model. First we collect data by interacting with the environment and save them to a replay memory file, and then fit and evaluate the Q network with the collected data. Repeat these two steps to improve the performance gradually.
 
 ### Dependencies:
 + [paddlepaddle==1.8.5](https://github.com/PaddlePaddle/Paddle)

benchmark/torch/ES/train.py

@@ -82,7 +82,7 @@ def step(self):
 
         while num_episodes < self.config['min_episodes_per_batch'] or \
                 num_timesteps < self.config['min_steps_per_batch']:
-            # setting the lastest params to the actor, getting the fitness and the noise seed sync.
+            # setting the latest params to the actor, getting the fitness and the noise seed sync.
             future_object_ids = [remote_actor.sample(self.latest_flat_weights) \
                 for remote_actor in self.remote_actors]
             results = [

benchmark/torch/a2c/train.py

@@ -77,7 +77,7 @@ def step(self):
         """
         1.setting latest_params to each actor model
         2.getting the sample data from all the actors synchronizely
-        3.traing the model with the sample data and the params is upgraded, and goto step 1
+        3.training the model with the sample data and the params is upgraded, and goto step 1
         """
 
         latest_params = self.agent.get_weights()

benchmark/torch/maddpg/train.py

@@ -124,7 +124,7 @@ def main():
         agents.append(agent)
 
     if args.restore:
-        # restore modle
+        # restore model
         for i in range(len(agents)):
             model_file = args.model_dir + '/agent_' + str(i)
             if not os.path.exists(model_file):

benchmark/torch/mappo/train.py

@@ -82,7 +82,7 @@ def main():
         buffers.append(bu)
 
     if args.restore:
-        # restore modle
+        # restore model
         for i in range(len(agents)):
             model_file = args.model_dir + '/' + args.env_name + '/agent_' + str(
                 i)

docs/basic_structure/agent.rst

@@ -14,13 +14,13 @@ Methods
 
 3. predict(self, obs)
 
-    Predict the action with current observation of the enviroment. Note that this function will only do the prediction and it doesn't try any exploration.
+    Predict the action with current observation of the environment. Note that this function will only do the prediction and it doesn't try any exploration.
     To explore in the action space, you should create your process in `sample` function below.
     Basically, this function is often used in test process.
 
 4. sample(self, obs)
 
-    Predict the action given current observation of the enviroment. 
+    Predict the action given current observation of the environment. 
     Additionaly, action will be added noise here to explore a new trajectory. 
     Basically, this function is often used in training process.
 

docs/locale/zh_CN/LC_MESSAGES/basic_structure.po

@@ -53,7 +53,7 @@ msgstr ""
 
 #: ../../basic_structure/agent.rst:17
 msgid ""
-"Predict the action with current observation of the enviroment. Note that "
+"Predict the action with current observation of the environment. Note that "
 "this function will only do the prediction and it doesn't try any "
 "exploration. To explore in the action space, you should create your "
 "process in `sample` function below. Basically, this function is often "
@@ -66,7 +66,7 @@ msgstr ""
 
 #: ../../basic_structure/agent.rst:23
 msgid ""
-"Predict the action given current observation of the enviroment. "
+"Predict the action given current observation of the environment. "
 "Additionaly, action will be added noise here to explore a new trajectory."
 " Basically, this function is often used in training process."
 msgstr ""

examples/Baselines/Halite_competition/paddle/rl_trainer/algorithm.py

@@ -103,7 +103,7 @@ def sample(self, obs):
         Return:
             value (paddle.tensor): predicted state value
             action (paddle.tensor): actions sampled from a distribution
-            action_log_probs (paddle.tensor): the log probabilites of action
+            action_log_probs (paddle.tensor): the log probabilities of action
         """
 
         batch_size = obs.shape[0]

examples/Baselines/Halite_competition/torch/rl_trainer/algorithm.py

@@ -107,7 +107,7 @@ def sample(self, obs):
         Return:
             value (torch.tensor): predicted state value
             action (torch.tensor): actions sampled from a distribution
-            action_log_probs (torch.tensor): the log probabilites of action
+            action_log_probs (torch.tensor): the log probabilities of action
         """
         with torch.no_grad():
             value = self.critic(obs)

examples/ES/train.py

@@ -84,7 +84,7 @@ def step(self):
 
         while num_episodes < self.config['min_episodes_per_batch'] or \
                 num_timesteps < self.config['min_steps_per_batch']:
-            # setting the lastest to the actors and get the fitness, noise seed sync.
+            # setting the latest to the actors and get the fitness, noise seed sync.
             future_object_ids = [remote_actor.sample(self.latest_flat_weights) \
                 for remote_actor in self.remote_actors]
             results = [

examples/MADDPG/train.py

@@ -124,7 +124,7 @@ def main():
         agents.append(agent)
 
     if args.restore:
-        # restore modle
+        # restore model
         for i in range(len(agents)):
             model_file = args.model_dir + '/agent_' + str(i)
             if not os.path.exists(model_file):

examples/NeurIPS2018-AI-for-Prosthetics-Challenge/simulator_server.py

@@ -113,7 +113,7 @@ def _new_ready_client(self):
 
     def Send(self, request, context):
         """ Implement Send function in SimulatorServicer
-        Everytime a request comming, will create a new thread to handle
+        Everytime a request coming, will create a new thread to handle
     """
         ident, obs, reward, done, info = request.id, request.observation, request.reward, request.done, request.info
         client = self.clients[ident]

examples/others/deepes.py

@@ -79,20 +79,20 @@ def reward_normalize(reward):
     for epcho in range(100):
         rewards = []
         noises = []
-        lastest_flat_weights = agent.get_flat_weights()
+        latest_flat_weights = agent.get_flat_weights()
 
         for episode in range(10):
             noise = np.random.randn(agent.weights_total_size)
             perturbation = noise * 0.05
 
-            agent.set_flat_weights(lastest_flat_weights + perturbation)
+            agent.set_flat_weights(latest_flat_weights + perturbation)
             ep_reward = evaluate(env, agent)
 
             noises.append(noise)
             rewards.append(ep_reward)
 
         normalized_rewards = reward_normalize(rewards)
-        agent.set_flat_weights(lastest_flat_weights)
+        agent.set_flat_weights(latest_flat_weights)
         agent.learn(normalized_rewards, noises)
         # evaluate
         if (epcho % 10) == 0:

parl/core/agent_base.py

@@ -54,7 +54,7 @@ def learn(self, *args, **kwargs):
         raise NotImplementedError
 
     def predict(self, *args, **kwargs):
-        """Predict the action when given the observation of the enviroment.
+        """Predict the action when given the observation of the environment.
 
         In general, this function is used in test process.
 
@@ -66,7 +66,7 @@ def predict(self, *args, **kwargs):
         raise NotImplementedError
 
     def sample(self, *args, **kwargs):
-        """Sample the action when given the observation of the enviroment.
+        """Sample the action when given the observation of the environment.
             
         In general, this function is used in train process.
 

parl/core/paddle/tests/policy_distribution_test_paddle.py

@@ -132,9 +132,9 @@ def test_kl(self):
         dist2 = self.get_dist(mean=mean, log_std=paddle.log(std2))
         kl = dist1.kl(dist2)
         single_kl_expect = np.log(2) - (3 / 8)
-        expect_ouput = paddle.ones_like(
+        expect_output = paddle.ones_like(
             kl) * self.num_actions * single_kl_expect
-        self.assertTrue(self.paddle_check_eq(kl, expect_ouput))
+        self.assertTrue(self.paddle_check_eq(kl, expect_output))
 
     def test_init_with_wrong_logits_shape(self):
         # input logits with wrong shape
@@ -169,7 +169,7 @@ def test_sample(self):
         self.assertGreaterEqual(paddle.max(sample_action), 0)
         self.assertLess(paddle.min(sample_action), self.num_actions)
 
-        # construct a logit to ouput determined class
+        # construct a logit to output determined class
         fool_logits = paddle.zeros(shape=(self.batch_size, self.num_actions))
         fool_logits[:, 0] = 9999999999
         set_random_seed(12)

parl/core/torch/tests/policy_distribution_test_torch.py

@@ -129,8 +129,8 @@ def test_kl(self):
         dist2 = self.get_dist(mean=mean, logstd=torch.log(std2))
         kl = dist1.kl(dist2)
         single_kl_expect = np.log(2) - (3 / 8)
-        expect_ouput = self.num_actions * single_kl_expect
-        self.assertTrue(self.torch_check_eq(expect_ouput, kl))
+        expect_output = self.num_actions * single_kl_expect
+        self.assertTrue(self.torch_check_eq(expect_output, kl))
 
     def test_init_with_wrong_logits_shape(self):
         # input logits with wrong shape
@@ -165,7 +165,7 @@ def test_sample(self):
         self.assertGreaterEqual(torch.max(sample_action), 0)
         self.assertLess(torch.min(sample_action), self.num_actions)
 
-        # construct a logit to ouput determined class
+        # construct a logit to output determined class
         fool_logits = torch.zeros(size=(self.batch_size, self.num_actions))
         fool_logits[:, 0] = 9999999999
         set_random_seed(12)

parl/env/continuous_wrappers.py

@@ -42,7 +42,7 @@ def reset(self, **kwargs):
     def step(self, model_output_act):
         """
         Args:
-            model_output_act(np.array): The values must be in in [-1, 1].
+            model_output_act(np.array): The values must be in [-1, 1].
         """
         assert np.all(((model_output_act<=1.0 + 1e-3), (model_output_act>=-1.0 - 1e-3))), \
             'the action should be in range [-1.0, 1.0]'

parl/remote/client.py

@@ -245,11 +245,11 @@ def check_env_consistency(self):
             master_pyarrow_version = to_str(message[4])
             if client_pyarrow_version != master_pyarrow_version:
                 if master_pyarrow_version == 'None':
-                    error_message = """"pyarrow" is provided in your current enviroment, however, it is not \
+                    error_message = """"pyarrow" is provided in your current environment, however, it is not \
 found in "master"'s environment. To use "pyarrow" for serialization, please install \
 "pyarrow={}" in "master"'s environment!""".format(client_pyarrow_version)
                 elif client_pyarrow_version == 'None':
-                    error_message = """"pyarrow" is provided in "master"'s enviroment, however, it is not \
+                    error_message = """"pyarrow" is provided in "master"'s environment, however, it is not \
 found in your current environment. To use "pyarrow" for serialization, please install \
 "pyarrow={}" in your current environment!""".format(master_pyarrow_version)
                 else:

parl/remote/exceptions.py

@@ -31,7 +31,7 @@ class RemoteError(Exception):
     """
 
     def __init__(self, func_name, error_info):
-        self.error_info = "[Error occured in xparl, when calling " +\
+        self.error_info = "[Error occurred in xparl, when calling " +\
             "function `{}`]:\n{}".format(func_name, error_info)
 
     def __str__(self):

parl/remote/templates/jobs.html

@@ -90,7 +90,7 @@ <h5 class="modal-title" id="log-modal-title">
           </div>
           <div class="modal-footer">
             <a style="position: relative; left: -160px; font-size: small;">
-              * Only the lastest 500 lines of the log are shown, <br />
+              * Only the latest 500 lines of the log are shown, <br />
               download the log file for the full log.
             </a>
             <button

parl/remote/worker.py

@@ -154,11 +154,11 @@ def check_env_consistency(self):
             master_pyarrow_version = to_str(message[4])
             if worker_pyarrow_version != master_pyarrow_version:
                 if master_pyarrow_version == 'None':
-                    error_message = """"pyarrow" is provided in your current enviroment, however, it is not \
+                    error_message = """"pyarrow" is provided in your current environment, however, it is not \
 found in "master"'s environment. To use "pyarrow" for serialization, please install \
 "pyarrow={}" in "master"'s environment!""".format(worker_pyarrow_version)
                 elif worker_pyarrow_version == 'None':
-                    error_message = """"pyarrow" is provided in "master"'s enviroment, however, it is not \
+                    error_message = """"pyarrow" is provided in "master"'s environment, however, it is not \
 found in your current environment. To use "pyarrow" for serialization, please install \
 "pyarrow={}" in your current environment!""".format(master_pyarrow_version)
                 else:

parl/utils/csv_logger.py

@@ -48,7 +48,7 @@ def __init__(self, output_file):
         self.keys_set = None
 
     def log_dict(self, result):
-        """Ouput result to the csv file.
+        """Output result to the csv file.
 
         Will create the header of the csv file automatically when the function is called for the first time.
         Ususally, the keys of the result should be the same every time you call the function.

parl/utils/path_utils.py

@@ -18,7 +18,7 @@
 
 
 def format_uniform_path(path):
-    """format the path to a new path which seperated by os.sep.
+    """format the path to a new path which separated by os.sep.
     """
     path = path.replace("//", os.sep)
     path = path.replace("/", os.sep)

parl/utils/tests/not_import_dl_framework_test.py

@@ -19,7 +19,7 @@
 
 class TestNotImportPaddle(unittest.TestCase):
     def test_import(self):
-        # setting this enviroment variable will not import deep learning framework
+        # setting this environment variable will not import deep learning framework
         os.environ['XPARL_igonre_core'] = 'true'
         import parl
         self.assertFalse('paddle' in sys.modules)

parl/utils/tests/not_import_dl_framework_test_torch.py

@@ -19,7 +19,7 @@
 
 class TestNotImportPaddle(unittest.TestCase):
     def test_import(self):
-        # setting this enviroment variable will not import deep learning framework
+        # setting this environment variable will not import deep learning framework
         os.environ['XPARL_igonre_core'] = 'true'
         import parl
         self.assertFalse('torch' in sys.modules)