From ea93665fcb6943ea4999ab8bd8faf30a8c2a2049 Mon Sep 17 00:00:00 2001
From: Guillaume Mollard <guillaume@iccluster028.iccluster.epfl.ch>
Date: Fri, 14 Jun 2019 10:05:42 +0200
Subject: [PATCH] corrected score metric
---
RLLib_training/RailEnvRLLibWrapper.py | 3 ++-
RLLib_training/train_experiment.py | 20 +++++++++++---------
2 files changed, 13 insertions(+), 10 deletions(-)
diff --git a/RLLib_training/RailEnvRLLibWrapper.py b/RLLib_training/RailEnvRLLibWrapper.py
index e36f383..8f7e1ec 100644
--- a/RLLib_training/RailEnvRLLibWrapper.py
+++ b/RLLib_training/RailEnvRLLibWrapper.py
@@ -20,7 +20,8 @@ class RailEnvRLLibWrapper(MultiAgentEnv):
self.predefined_env = False
if config['rail_generator'] == "complex_rail_generator":
- self.rail_generator = complex_rail_generator(nr_start_goal=config['number_of_agents'], min_dist=5,
+ self.rail_generator = complex_rail_generator(nr_start_goal=config['number_of_agents'],
+ min_dist=config['min_dist'],
nr_extra=config['nr_extra'],
seed=config['seed'] * (1 + vector_index))
elif config['rail_generator'] == "random_rail_generator":
diff --git a/RLLib_training/train_experiment.py b/RLLib_training/train_experiment.py
index b674d40..cc8debe 100644
--- a/RLLib_training/train_experiment.py
+++ b/RLLib_training/train_experiment.py
@@ -62,8 +62,8 @@ def on_episode_end(info):
episode = info['episode']
score = 0
for k, v in episode._agent_reward_history.items():
- score += np.mean(v)
- score /= (len(episode._agent_reward_history) * 1.5 * episode.horizon)
+ score += np.sum(v)
+ score /= (len(episode._agent_reward_history) * 3 * episode.horizon)
episode.custom_metrics["score"] = score
def train(config, reporter):
@@ -80,6 +80,7 @@ def train(config, reporter):
"number_of_agents": config['n_agents'],
"seed": config['seed'],
"obs_builder": config['obs_builder'],
+ "min_dist": config['min_dist'],
# "predictor": config["predictor"],
"step_memory": config["step_memory"]}
@@ -154,9 +155,9 @@ def train(config, reporter):
trainer_config["horizon"] = 1.5 * (config['map_width'] + config['map_height'])#config['horizon']
trainer_config["num_workers"] = 0
- trainer_config["num_cpus_per_worker"] = 4
- trainer_config["num_gpus"] = 0.2
- trainer_config["num_gpus_per_worker"] = 0.2
+ trainer_config["num_cpus_per_worker"] = 7
+ trainer_config["num_gpus"] = 0.0
+ trainer_config["num_gpus_per_worker"] = 0.0
trainer_config["num_cpus_for_driver"] = 1
trainer_config["num_envs_per_worker"] = 1
trainer_config['entropy_coeff'] = config['entropy_coeff']
@@ -203,7 +204,7 @@ def train(config, reporter):
def run_experiment(name, num_iterations, n_agents, hidden_sizes, save_every,
map_width, map_height, horizon, policy_folder_name, local_dir, obs_builder,
entropy_coeff, seed, conv_model, rail_generator, nr_extra, kl_coeff, lambda_gae,
- step_memory):
+ step_memory, min_dist):
tune.run(
train,
name=name,
@@ -224,11 +225,12 @@ def run_experiment(name, num_iterations, n_agents, hidden_sizes, save_every,
"nr_extra": nr_extra,
"kl_coeff": kl_coeff,
"lambda_gae": lambda_gae,
+ "min_dist": min_dist,
# "predictor": predictor,
"step_memory": step_memory
},
resources_per_trial={
- "cpu": 2,
+ "cpu": 8,
"gpu": 0
},
verbose=2,
@@ -240,7 +242,7 @@ if __name__ == '__main__':
gin.external_configurable(tune.grid_search)
# with path('RLLib_training.experiment_configs.n_agents_experiment', 'config.gin') as f:
# gin.parse_config_file(f)
- gin.parse_config_file('/home/guillaume/EPFL/Master_Thesis/flatland/baselines/RLLib_training/experiment_configs/predictions_test/config.gin')
- dir = '/home/guillaume/EPFL/Master_Thesis/flatland/baselines/RLLib_training/experiment_configs/predictions_test'
+ gin.parse_config_file('/home/guillaume/flatland/baselines/RLLib_training/experiment_configs/score_metric_test/config.gin')
+ dir = '/home/guillaume/flatland/baselines/RLLib_training/experiment_configs/score_metric_test'
# dir = os.path.join(__file_dirname__, 'experiment_configs', 'experiment_agent_memory')
run_experiment(local_dir=dir)
--
GitLab