wassname/ray
1
0
Fork 0
mirror of https://github.com/wassname/ray.git synced 2026-07-18 12:40:56 +08:00
Code Issues Packages Projects Releases Wiki Activity

Clarify training intensity configuration docstring (#9244) (#9306)

Browse Source
This commit is contained in:
Piotr Januszewski
2020-07-05 20:07:27 -07:00
committed by GitHub
parent 6f3d993681
commit 155cc81e40
5 changed files with 19 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files
rllib/agents/ddpg/ddpg.py
+4 -3
View File
@@ -106,9 +106,10 @@ DEFAULT_CONFIG = with_common_config({
"prioritized_replay_eps": 1e-6,
# Whether to LZ4 compress observations
"compress_observations": False,
# If set, this will fix the ratio of sampled to replayed timesteps.
# Otherwise, replay will proceed at the native ratio determined by
# (train_batch_size / rollout_fragment_length).
# If set, this will fix the ratio of replayed from a buffer and learned on
# timesteps to sampled from an environment and stored in the replay buffer
# timesteps. Otherwise, the replay will proceed at the native ratio
# determined by (train_batch_size / rollout_fragment_length).
"training_intensity": None,
# === Optimization ===
rllib/agents/dqn/apex.py
+3 -2
View File
@@ -41,8 +41,9 @@ APEX_DEFAULT_CONFIG = merge_dicts(
"exploration_config": {"type": "PerWorkerEpsilonGreedy"},
"worker_side_prioritization": True,
"min_iter_time_s": 30,
# If set, this will fix the ratio of sampled to replayed timesteps.
# Otherwise, replay will proceed as fast as possible.
# If set, this will fix the ratio of replayed from a buffer and learned
# on timesteps to sampled from an environment and stored in the replay
# buffer timesteps. Otherwise, replay will proceed as fast as possible.
"training_intensity": None,
},
)
rllib/agents/dqn/dqn.py
+4 -3
View File
@@ -85,9 +85,10 @@ DEFAULT_CONFIG = with_common_config({
"compress_observations": False,
# Callback to run before learning on a multi-agent batch of experiences.
"before_learn_on_batch": None,
# If set, this will fix the ratio of sampled to replayed timesteps.
# Otherwise, replay will proceed at the native ratio determined by
# (train_batch_size / rollout_fragment_length).
# If set, this will fix the ratio of replayed from a buffer and learned on
# timesteps to sampled from an environment and stored in the replay buffer
# timesteps. Otherwise, the replay will proceed at the native ratio
# determined by (train_batch_size / rollout_fragment_length).
"training_intensity": None,
# === Optimization ===
rllib/agents/sac/sac.py
+4 -3
View File
@@ -68,9 +68,10 @@ DEFAULT_CONFIG = with_common_config({
"final_prioritized_replay_beta": 0.4,
# Whether to LZ4 compress observations
"compress_observations": False,
# If set, this will fix the ratio of sampled to replayed timesteps.
# Otherwise, replay will proceed at the native ratio determined by
# (train_batch_size / rollout_fragment_length).
# If set, this will fix the ratio of replayed from a buffer and learned on
# timesteps to sampled from an environment and stored in the replay buffer
# timesteps. Otherwise, the replay will proceed at the native ratio
# determined by (train_batch_size / rollout_fragment_length).
"training_intensity": None,
# === Optimization ===
rllib/contrib/maddpg/maddpg.py
+4 -3
View File
@@ -70,9 +70,10 @@ DEFAULT_CONFIG = with_common_config({
# Observation compression. Note that compression makes simulation slow in
# MPE.
"compress_observations": False,
# If set, this will fix the ratio of sampled to replayed timesteps.
# Otherwise, replay will proceed at the native ratio determined by
# (train_batch_size / rollout_fragment_length).
# If set, this will fix the ratio of replayed from a buffer and learned on
# timesteps to sampled from an environment and stored in the replay buffer
# timesteps. Otherwise, the replay will proceed at the native ratio
# determined by (train_batch_size / rollout_fragment_length).
"training_intensity": None,
# Force lockstep replay mode for MADDPG.
"multiagent": merge_dicts(COMMON_CONFIG["multiagent"], {