huangfu/InteractionImitation
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Generate videos of expert data

Browse Source
This commit is contained in:
ebuehrle
2022-03-04 16:07:15 +01:00
parent 1e70303c57
commit 575e299fc8
3 changed files with 6 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
eval_experiments.py
View File

@@ -11,7 +11,7 @@ def main(method:str='expert', folder:str=None, locations=[(0,0)], skip_running=F
exclude_keys_from_policy_kwargs = {'learning_rate', 'learning_rate_decay', 'clip_ratio', 'iterations_per_epoch', 'option'}
policy_kwargs = {}
if method in ['expert', 'idm']:
if method in ['expert', 'expert_agent', 'idm']:
env, env_kwargs ='NRasterizedRouteIncrementingAgent', {}
elif method in ['bc','gail']:
env='NormalizedContinuousEvalEnv'

2
generate_videos.sh
View File

@@ -3,6 +3,7 @@
# Experiment A
python -m eval_experiments
python -m eval_experiments --method expert_agent --save_videos
python -m eval_experiments --method idm --save_videos
python -m eval_experiments --method bc --folder='test_policies/bc/expA' --save_videos
python -m eval_experiments --method gail --folder='test_policies/gail/expA' --save_videos
@@ -11,6 +12,7 @@ python -m eval_experiments --method shail --folder='test_policies/shail/expA' --
# Experiment B
python -m eval_experiments --locations='[(0,4)]'
python -m eval_experiments --method expert_agent --locations='[(0,4)]' --save_videos
python -m eval_experiments --method idm --locations='[(0,4)]' --save_videos
python -m eval_experiments --method bc --folder='test_policies/bc/expB' --locations='[(0,4)]' --save_videos
python -m eval_experiments --method gail --folder='test_policies/gail/expB' --locations='[(0,4)]' --save_videos

3
src/eval_main.py
View File

@@ -5,6 +5,7 @@ import intersim
from intersim.envs import Intersimple
from stable_baselines3.common.base_class import BaseAlgorithm
from src.baselines import IDMRulePolicy
from src.data.expert import NormalizedIntersimpleExpert
from src.evaluation import IntersimpleEvaluation
import src.gail.options as options_envs
from src.evaluation.metrics import divergence, visualize_distribution, rwse
@@ -40,6 +41,8 @@ def load_policy(method:str,
ml = torch.device('cpu') if not torch.cuda.is_available() else None
if method == 'idm':
policy = IDMRulePolicy(env, **policy_kwargs)
elif method == 'expert_agent':
policy = NormalizedIntersimpleExpert(env, **policy_kwargs)
elif method == 'bc':
policy = SetPolicy(env.action_space.shape[-1], **policy_kwargs)
policy.load_state_dict(torch.load(policy_file, map_location=ml))