Develop timeout (#5991)

* Added timeout cli.

* Updated docs.

* Updated mlagents learn cli test.

com.unity.ml-agents/CHANGELOG.md

@@ -9,20 +9,21 @@ and this project adheres to
 ## [Unreleased]
 ### Major Changes
 #### com.unity.ml-agents / com.unity.ml-agents.extensions (C#)
-- Updated to PyTorch 1.13.1
-- Deprecated support for Python 3.8.x and 3.9.x
 - Upgraded ML-Agents to Sentis 1.2.0-exp.2 (#)
 - The minimum supported Unity version was updated to 2022.3. (#)
 - Added batched raycast sensor option. (#)
 
 #### ml-agents / ml-agents-envs
+- Updated to PyTorch 1.13.1
+- Deprecated support for Python 3.8.x and 3.9.x
 
 ### Minor Changes
 #### com.unity.ml-agents / com.unity.ml-agents.extensions (C#)
 - Added DecisionStep parameter to DecisionRequester (#)
   - This will allow the staggering of execution timing when using multi-agents, leading to more stable performance.
 
 #### ml-agents / ml-agents-envs
+- Added timeout cli and yaml config file support for specifying environment timeout.
 - Added training config feature to evenly distribute checkpoints throughout training. (#5842)
 - Updated training area replicator to add a condition to only replicate training areas when running a build. (#5842)
 

docs/Training-ML-Agents.md

@@ -213,6 +213,7 @@ env_settings:
   env_args: null
   base_port: 5005
   num_envs: 1
+  timeout_wait: 10
   seed: -1
   max_lifetime_restarts: 10
   restarts_rate_limit_n: 1

ml-agents/mlagents/trainers/cli_utils.py

@@ -233,6 +233,13 @@ def _create_parser() -> argparse.ArgumentParser:
         help="Results base directory",
     )
 
+    argparser.add_argument(
+        "--timeout-wait",
+        default=60,
+        help="The period of time to wait on a Unity environment to startup for training.",
+        action=DetectDefault,
+    )
+
     eng_conf = argparser.add_argument_group(title="Engine Configuration")
     eng_conf.add_argument(
         "--width",

ml-agents/mlagents/trainers/learn.py

@@ -101,6 +101,7 @@ def run_training(run_seed: int, options: RunOptions, num_areas: int) -> None:
             engine_settings.no_graphics,
             run_seed,
             num_areas,
+            env_settings.timeout_wait,
             port,
             env_settings.env_args,
             os.path.abspath(run_logs_dir),  # Unity environment requires absolute path
@@ -175,6 +176,7 @@ def create_environment_factory(
     no_graphics: bool,
     seed: int,
     num_areas: int,
+    timeout_wait: int,
     start_port: Optional[int],
     env_args: Optional[List[str]],
     log_folder: str,
@@ -194,6 +196,7 @@ def create_unity_environment(
             additional_args=env_args,
             side_channels=side_channels,
             log_folder=log_folder,
+            timeout_wait=timeout_wait,
         )
 
     return create_unity_environment

ml-agents/mlagents/trainers/settings.py

@@ -817,6 +817,7 @@ class EnvironmentSettings:
     base_port: int = parser.get_default("base_port")
     num_envs: int = attr.ib(default=parser.get_default("num_envs"))
     num_areas: int = attr.ib(default=parser.get_default("num_areas"))
+    timeout_wait: int = attr.ib(default=parser.get_default("timeout_wait"))
     seed: int = parser.get_default("seed")
     max_lifetime_restarts: int = parser.get_default("max_lifetime_restarts")
     restarts_rate_limit_n: int = parser.get_default("restarts_rate_limit_n")

ml-agents/mlagents/trainers/tests/test_learn.py

@@ -105,6 +105,7 @@ def test_bad_env_path():
             no_graphics=True,
             seed=-1,
             num_areas=1,
+            timeout_wait=1,
             start_port=8000,
             env_args=None,
             log_folder="results/log_folder",