wassname/ray
1
0
Fork 0
mirror of https://github.com/wassname/ray.git synced 2026-08-09 12:20:09 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix old exploration configs. (#7240)

Browse Source
This commit is contained in:
Sven Mika
2020-02-20 08:39:16 -08:00
committed by GitHub
parent 1ae7c03e86
commit 6043ce710d
9 changed files with 10 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files
rllib/agents/dqn/apex.py
+1 -1
View File
@@ -22,7 +22,7 @@ APEX_DEFAULT_CONFIG = merge_dicts(
"sample_batch_size": 50,
"target_network_update_freq": 500000,
"timesteps_per_iteration": 25000,
"exploration": {"type": "PerWorkerEpsilonGreedy"},
"exploration_config": {"type": "PerWorkerEpsilonGreedy"},
"worker_side_prioritization": True,
"min_iter_time_s": 30,
},
rllib/tests/test_ignore_worker_failure.py
+1 -1
View File
@@ -72,7 +72,7 @@ class IgnoresWorkerFailure(unittest.TestCase):
"timesteps_per_iteration": 1000,
"num_gpus": 0,
"min_iter_time_s": 1,
"exploration": False,
"explore": False,
"learning_starts": 1000,
"target_network_update_freq": 100,
"optimizer": {
rllib/tuned_examples/atari-apex.yaml
+1 -1
View File
@@ -17,7 +17,7 @@ apex:
adam_epsilon: .00015
hiddens: [512]
buffer_size: 1000000
exploration:
exploration_config:
final_epsilon: 0.01
epsilon_timesteps: 200000
prioritized_replay_alpha: 0.5
rllib/tuned_examples/atari-dist-dqn.yaml
+1 -1
View File
@@ -21,7 +21,7 @@ atari-dist-dqn:
buffer_size: 1000000
sample_batch_size: 4
train_batch_size: 32
exploration:
exploration_config:
epsilon_timesteps: 200000
final_epsilon: 0.01
prioritized_replay_alpha: 0.5
rllib/tuned_examples/atari-dqn.yaml
+1 -1
View File
@@ -23,7 +23,7 @@ atari-basic-dqn:
buffer_size: 1000000
sample_batch_size: 4
train_batch_size: 32
exploration:
exploration_config:
epsilon_timesteps: 200000
final_epsilon: 0.01
prioritized_replay_alpha: 0.5
rllib/tuned_examples/atari-duel-ddqn.yaml
+1 -1
View File
@@ -23,7 +23,7 @@ dueling-ddqn:
buffer_size: 1000000
sample_batch_size: 4
train_batch_size: 32
exploration:
exploration_config:
epsilon_timesteps: 200000
final_epsilon: 0.01
prioritized_replay_alpha: 0.5
rllib/tuned_examples/compact-regression-test.yaml
+2 -2
View File
@@ -85,7 +85,7 @@ apex:
adam_epsilon: .00015
hiddens: [512]
buffer_size: 1000000
exploration:
exploration_config:
epsilon_timesteps: 200000
final_epsilon: 0.01
prioritized_replay_alpha: 0.5
@@ -135,7 +135,7 @@ atari-basic-dqn:
buffer_size: 1000000
sample_batch_size: 4
train_batch_size: 32
exploration:
exploration_config:
epsilon_timesteps: 200000
final_epsilon: 0.01
prioritized_replay_alpha: 0.5
rllib/tuned_examples/pong-dqn.yaml
+1 -1
View File
@@ -13,7 +13,7 @@ pong-deterministic-dqn:
buffer_size: 50000
sample_batch_size: 4
train_batch_size: 32
exploration:
exploration_config:
epsilon_timesteps: 200000
final_epsilon: .01
model:
rllib/tuned_examples/pong-rainbow.yaml
+1 -1
View File
@@ -13,7 +13,7 @@ pong-deterministic-rainbow:
buffer_size: 50000
sample_batch_size: 4
train_batch_size: 32
exploration:
exploration_config:
epsilon_timesteps: 2
final_epsilon: 0.0
target_network_update_freq: 500