Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[RLlib] Fix/remove some CI tests and many_ppo release test. #47686

Merged
merged 2 commits into from
Sep 16, 2024
+10 −53
Merged

[RLlib] Fix/remove some CI tests and many_ppo release test. #47686

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
c68373d
wip
sven1977 Sep 16, 2024
117ebc5
wip
sven1977 Sep 16, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 3 additions & 3 deletions release/long_running_tests/workloads/many_ppo.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,6 +2,7 @@
# This covers https://github.com/ray-project/ray/pull/12148

import ray
from ray.rllib.utils.metrics import NUM_ENV_STEPS_SAMPLED_LIFETIME
from ray.tune import run_experiments
from ray.tune.utils.release_test_util import ProgressCallback
from ray._private.test_utils import monitor_memory_usage
Expand Down Expand Up @@ -34,13 +35,12 @@
"env": "CartPole-v0",
"num_samples": 10000,
"config": {
"framework": "torch",
"num_workers": 7,
"num_env_runners": 7,
"num_gpus": 0,
"num_sgd_iter": 1,
},
"stop": {
"timesteps_total": 1,
NUM_ENV_STEPS_SAMPLED_LIFETIME: 1,
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Does this mean, we only make quick tests if the class can be built and the algorithm runs?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yeah, I guess 🤷 . I think this is a conceptual test for tune to be able to handle large amounts of samples, and not a stress test.

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for clarifying this :)

},
}
},
Expand Down
46 changes: 0 additions & 46 deletions rllib/BUILD
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -319,52 +319,6 @@ py_test(
args = ["--dir=tuned_examples/appo"]
)

# Tests against crashing or hanging environments.
# Single-agent: Crash only.
#@OldAPIStack
py_test(
name = "learning_tests_cartpole_crashing_appo_old_api_stack",
main = "tests/run_regression_tests.py",
tags = ["team:rllib", "exclusive", "learning_tests", "learning_tests_discrete", "crashing_cartpole", "torch_only"],
size = "large",
srcs = ["tests/run_regression_tests.py"],
data = ["tuned_examples/appo/cartpole-crashing-recreate-workers-appo.py"],
args = ["--dir=tuned_examples/appo", "--num-cpus=6"]
)
# Single-agent: Crash and stall.
#@OldAPIStack
py_test(
name = "learning_tests_cartpole_crashing_and_stalling_appo_old_api_stack",
main = "tests/run_regression_tests.py",
tags = ["team:rllib", "exclusive", "learning_tests", "learning_tests_discrete", "crashing_cartpole", "torch_only"],
size = "large",
srcs = ["tests/run_regression_tests.py"],
data = ["tuned_examples/appo/cartpole-crashing-and-stalling-recreate-workers-appo.py"],
args = ["--dir=tuned_examples/appo", "--num-cpus=6"]
)
# Multi-agent: Crash only.
#@OldAPIStack
py_test(
name = "learning_tests_multi_agent_cartpole_crashing_appo_old_api_stack",
main = "tests/run_regression_tests.py",
tags = ["team:rllib", "exclusive", "learning_tests", "learning_tests_discrete", "crashing_cartpole", "torch_only"],
size = "large",
srcs = ["tests/run_regression_tests.py"],
data = ["tuned_examples/appo/multi-agent-cartpole-crashing-recreate-workers-appo.py"],
args = ["--dir=tuned_examples/appo", "--num-cpus=6"]
)
# Multi-agent: Crash and stall.
#@OldAPIStack
py_test(
name = "learning_tests_multi_agent_cartpole_crashing_and_stalling_appo_old_api_stack",
main = "tests/run_regression_tests.py",
tags = ["team:rllib", "exclusive", "learning_tests", "learning_tests_discrete", "crashing_cartpole", "torch_only"],
size = "large",
srcs = ["tests/run_regression_tests.py"],
data = ["tuned_examples/appo/multi-agent-cartpole-crashing-and-stalling-recreate-workers-appo.py"],
args = ["--dir=tuned_examples/appo", "--num-cpus=6"]
)

# BC
# CartPole
py_test(
Expand Down
2 changes: 1 addition & 1 deletion rllib/tuned_examples/ppo/atari_ppo.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -31,7 +31,7 @@ def _make_learner_connector(input_observation_space, input_action_space):
# We would like our frame stacking connector to do this job.
def _env_creator(cfg):
return wrap_atari_for_new_api_stack(
gym.make(args.env, **cfg),
gym.make(args.env, **cfg, render_mode="rgb_array"),
# Perform frame-stacking through ConnectorV2 API.
framestack=None,
)
Expand Down
9 changes: 6 additions & 3 deletions rllib/utils/test_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,7 +30,7 @@
import ray
from ray import air, tune
from ray.air.constants import TRAINING_ITERATION
from ray.air.integrations.wandb import WandbLoggerCallback
from ray.air.integrations.wandb import WandbLoggerCallback, WANDB_ENV_VAR
from ray.rllib.common import SupportedFileType
from ray.rllib.env.wrappers.atari_wrappers import is_atari, wrap_deepmind
from ray.rllib.train import load_experiments_from_file
Expand Down Expand Up @@ -1462,13 +1462,16 @@ def run_rllib_example_script_experiment(

# Log results using WandB.
tune_callbacks = tune_callbacks or []
if hasattr(args, "wandb_key") and args.wandb_key is not None:
if hasattr(args, "wandb_key") and (
args.wandb_key is not None or os.environ[WANDB_ENV_VAR]
):
wandb_key = args.wandb_key or os.environ[WANDB_ENV_VAR]
project = args.wandb_project or (
args.algo.lower() + "-" + re.sub("\\W+", "-", str(config.env).lower())
)
tune_callbacks.append(
WandbLoggerCallback(
api_key=args.wandb_key,
api_key=wandb_key,
project=project,
upload_checkpoints=True,
**({"name": args.wandb_run_name} if args.wandb_run_name else {}),
Expand Down
Loading