wassname/ray
1
0
Fork 0
mirror of https://github.com/wassname/ray.git synced 2026-07-17 11:32:33 +08:00
Code Issues Packages Projects Releases Wiki Activity

[rllib] more user-friendly Optimizer signature + compute_apply (#2335)

Browse Source 
* Move signature of optimizers

* fix

* expose compute_apply for policy_graphs

* dictionaries and such

* test for multiagent
This commit is contained in:
Richard Liaw
2018-07-07 12:08:49 -07:00
committed by GitHub
parent e3534c46df
commit e32aed8717
10 changed files with 35 additions and 21 deletions
Download Patch File Download Diff File Expand all files Collapse all files
python/ray/rllib/agents/a3c/a3c.py
+2 -2
View File
@@ -94,8 +94,8 @@ class A3CAgent(Agent):
self.env_creator, policy_cls, self.config["num_workers"],
{"num_gpus": 1 if self.config["use_gpu_for_workers"] else 0})
self.optimizer = AsyncGradientsOptimizer(
self.config["optimizer"], self.local_evaluator,
self.remote_evaluators)
self.local_evaluator, self.remote_evaluators,
self.config["optimizer"])
def _train(self):
self.optimizer.step()
python/ray/rllib/agents/bc/bc.py
+2 -2
View File
@@ -72,8 +72,8 @@ class BCAgent(Agent):
remote_cls.remote(self.env_creator, self.config, self.logdir)
for _ in range(self.config["num_workers"])]
self.optimizer = AsyncGradientsOptimizer(
self.config["optimizer"], self.local_evaluator,
self.remote_evaluators)
self.local_evaluator, self.remote_evaluators,
self.config["optimizer"])
def _train(self):
self.optimizer.step()
python/ray/rllib/agents/dqn/dqn.py
+2 -2
View File
@@ -136,8 +136,8 @@ class DQNAgent(Agent):
{"num_cpus": self.config["num_cpus_per_worker"],
"num_gpus": self.config["num_gpus_per_worker"]})
self.optimizer = getattr(optimizers, self.config["optimizer_class"])(
self.config["optimizer"], self.local_evaluator,
self.remote_evaluators)
self.local_evaluator, self.remote_evaluators,
self.config["optimizer"])
self.last_target_update_ts = 0
self.num_target_updates = 0
python/ray/rllib/agents/pg/pg.py
+2 -2
View File
@@ -45,8 +45,8 @@ class PGAgent(Agent):
self.remote_evaluators = self.make_remote_evaluators(
self.env_creator, PGPolicyGraph, self.config["num_workers"], {})
self.optimizer = SyncSamplesOptimizer(
self.config["optimizer"], self.local_evaluator,
self.remote_evaluators)
self.local_evaluator, self.remote_evaluators,
self.config["optimizer"])
def _train(self):
self.optimizer.step()
python/ray/rllib/agents/ppo/ppo.py
+2 -2
View File
@@ -74,11 +74,11 @@ class PPOAgent(Agent):
{"num_cpus": self.config["num_cpus_per_worker"],
"num_gpus": self.config["num_gpus_per_worker"]})
self.optimizer = LocalMultiGPUOptimizer(
self.local_evaluator, self.remote_evaluators,
{"sgd_batch_size": self.config["sgd_batchsize"],
"sgd_stepsize": self.config["sgd_stepsize"],
"num_sgd_iter": self.config["num_sgd_iter"],
"timesteps_per_batch": self.config["timesteps_per_batch"]},
self.local_evaluator, self.remote_evaluators)
"timesteps_per_batch": self.config["timesteps_per_batch"]})
def _train(self):
def postprocess_samples(batch):