wassname/ray
1
0
Fork 0
mirror of https://github.com/wassname/ray.git synced 2026-07-20 12:40:20 +08:00
Code Issues Packages Projects Releases Wiki Activity

[rllib] Remove uses of sgd_stepsize => lr (#3667)

Browse Source 
* lr

* Update example-evolution-strategies.rst
This commit is contained in:
Eric Liang
2019-01-01 12:01:27 +08:00
committed by GitHub
parent 93d54110f8
commit b8a9e3f106
2 changed files with 5 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files
doc/source/example-evolution-strategies.rst
+1 -1
View File
@@ -28,7 +28,7 @@ To train a policy on a cluster (e.g., using 900 workers), run the following.
--env=Humanoid-v1 \
--run=ES \
--redis-address=<redis-address> \
--config='{"num_workers": 900, "episodes_per_batch": 10000, "timesteps_per_batch": 100000}'
--config='{"num_workers": 900, "episodes_per_batch": 10000, "train_batch_size": 100000}'
At the heart of this example, we define a ``Worker`` class. These workers have
a method ``do_rollouts``, which will be used to perform simulate randomly
doc/source/rllib-training.rst
+4 -4
View File
@@ -156,7 +156,7 @@ All RLlib agents are compatible with the `Tune API <tune-usage.html>`__. This en
"config": {
"num_gpus": 0,
"num_workers": 1,
"sgd_stepsize": tune.grid_search([0.01, 0.001, 0.0001]),
"lr": tune.grid_search([0.01, 0.001, 0.0001]),
},
},
})
@@ -170,10 +170,10 @@ Tune will schedule the trials to run in parallel on your Ray cluster:
Resources requested: 4/4 CPUs, 0/0 GPUs
Result logdir: ~/ray_results/my_experiment
PENDING trials:
- PPO_CartPole-v0_2_sgd_stepsize=0.0001: PENDING
- PPO_CartPole-v0_2_lr=0.0001: PENDING
RUNNING trials:
- PPO_CartPole-v0_0_sgd_stepsize=0.01: RUNNING [pid=21940], 16 s, 4013 ts, 22 rew
- PPO_CartPole-v0_1_sgd_stepsize=0.001: RUNNING [pid=21942], 27 s, 8111 ts, 54.7 rew
- PPO_CartPole-v0_0_lr=0.01: RUNNING [pid=21940], 16 s, 4013 ts, 22 rew
- PPO_CartPole-v0_1_lr=0.001: RUNNING [pid=21942], 27 s, 8111 ts, 54.7 rew
Accessing Policy State
~~~~~~~~~~~~~~~~~~~~~~