def test_transfer_learning(tmpdir):
"""Transfer learning smoke test.
Saves a dummy AIRL test reward, then loads it for transfer learning.
"""
log_dir_train = osp.join(tmpdir, "train")
run = train_ex.run(
named_configs=["cartpole", "airl", "fast"],
config_updates=dict(
rollout_path=
"tests/data/expert_models/cartpole_0/rollouts/final.pkl",
log_dir=log_dir_train,
),
)
assert run.status == "COMPLETED"
_check_train_ex_result(run.result)
_check_rollout_stats(run.result["imit_stats"])
log_dir_data = osp.join(tmpdir, "expert_demos")
discrim_path = osp.join(log_dir_train, "checkpoints", "final", "discrim")
run = expert_demos_ex.run(
named_configs=["cartpole", "fast"],
config_updates=dict(
log_dir=log_dir_data,
reward_type="DiscrimNet",
reward_path=discrim_path,
),
)
assert run.status == "COMPLETED"
_check_rollout_stats(run.result)
def test_transfer_learning():
"""Transfer learning smoke test.
Save a dummy AIRL test reward, then load it for transfer learning."""
with tempfile.TemporaryDirectory(prefix='imitation-transfer',
) as tmpdir:
log_dir_train = osp.join(tmpdir, "train")
run = train_ex.run(
named_configs=['cartpole', 'airl', 'fast'],
config_updates=dict(
rollout_glob="tests/data/rollouts/CartPole*.pkl",
log_dir=log_dir_train,
),
)
assert run.status == 'COMPLETED'
log_dir_data = osp.join(tmpdir, "expert_demos")
discrim_path = osp.join(log_dir_train, "checkpoints", "final", "discrim")
run = expert_demos_ex.run(
named_configs=['cartpole', 'fast'],
config_updates=dict(
log_dir=log_dir_data,
reward_type='DiscrimNetAIRL',
reward_path=discrim_path,
),
)
assert run.status == 'COMPLETED'
def test_train_adversarial(tmpdir):
"""Smoke test for imitation.scripts.train_adversarial."""
named_configs = ["cartpole", "gail", "fast"]
config_updates = {
"log_root": tmpdir,
"rollout_path":
"tests/data/expert_models/cartpole_0/rollouts/final.pkl",
"init_tensorboard": True,
"plot_interval": 1,
"extra_episode_data_interval": 1,
}
run = train_ex.run(
named_configs=named_configs,
config_updates=config_updates,
)
assert run.status == "COMPLETED"
_check_train_ex_result(run.result)
def test_train_adversarial(tmpdir):
"""Smoke test for imitation.scripts.train_adversarial"""
named_configs = ['cartpole', 'gail', 'fast']
config_updates = {
'log_root': tmpdir,
'rollout_path':
"tests/data/expert_models/cartpole_0/rollouts/final.pkl",
'init_tensorboard': True,
'plot_interval': 1,
'extra_episode_data_interval': 1,
}
run = train_ex.run(
named_configs=named_configs,
config_updates=config_updates,
)
assert run.status == 'COMPLETED'
_check_train_ex_result(run.result)
def test_train_adversarial():
"""Smoke test for imitation.scripts.train_adversarial"""
with tempfile.TemporaryDirectory(prefix='imitation-train',
) as tmpdir:
config_updates = {
'init_trainer_kwargs': {
# Rollouts are small, decrease size of buffer to avoid warning
'trainer_kwargs': {
'n_disc_samples_per_buffer': 50,
},
},
'log_root': tmpdir,
'rollout_glob': "tests/data/rollouts/CartPole*.pkl",
}
run = train_ex.run(
named_configs=['cartpole', 'gail', 'fast'],
config_updates=config_updates,
)
assert run.status == 'COMPLETED'
def test_train_adversarial(tmpdir):
"""Smoke test for imitation.scripts.train_adversarial"""
named_configs = ['cartpole', 'gail', 'fast', 'plots']
config_updates = {
'init_trainer_kwargs': {
# Rollouts are small, decrease size of buffer to avoid warning
'trainer_kwargs': {
'n_disc_samples_per_buffer': 50,
},
},
'log_root': tmpdir,
'rollout_path':
"tests/data/expert_models/cartpole_0/rollouts/final.pkl",
'init_tensorboard': True,
}
run = train_ex.run(
named_configs=named_configs,
config_updates=config_updates,
)
assert run.status == 'COMPLETED'
_check_train_ex_result(run.result)
def test_analyze_imitation(tmpdir: str, run_names: List[str]):
sacred_logs_dir = tmpdir
# Generate sacred logs (other logs are put in separate tmpdir for deletion).
for i, run_name in enumerate(run_names):
with tempfile.TemporaryDirectory(prefix="junk") as junkdir:
rollout_path = "tests/data/expert_models/cartpole_0/rollouts/final.pkl"
run = train_ex.run(
named_configs=["fast", "cartpole"],
config_updates=dict(
rollout_path=rollout_path,
log_dir=junkdir,
checkpoint_interval=-1,
),
options={
"--name": run_name,
"--file_storage": sacred_logs_dir
},
)
assert run.status == "COMPLETED"
# Check that analyze script finds the correct number of logs.
def check(run_name: Optional[str], count: int) -> None:
run = analysis_ex.run(
command_name="analyze_imitation",
config_updates=dict(
source_dir=sacred_logs_dir,
run_name=run_name,
csv_output_path=osp.join(tmpdir, "analysis.csv"),
verbose=True,
),
)
assert run.status == "COMPLETED"
df = pd.DataFrame(run.result)
assert df.shape[0] == count
for run_name, count in Counter(run_names).items():
check(run_name, count)
check(None, len(run_names)) # Check total number of logs.