hiro/ray
1
0
Fork 0
mirror of https://github.com/vale981/ray synced 2025-03-05 10:01:43 -05:00
Code Issues Projects Releases Packages Wiki Activity Actions

[RLlib] Add DQN SoftQ learning test case. (#12712)

Browse source
This commit is contained in:
Sven Mika 2020-12-10 14:55:19 +01:00 • committed by GitHub
parent e3b5deb741
commit deb33bce84
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
2 changed files with 18 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
rllib/BUILD
View file

@ -172,6 +172,7 @@ py_test(
data = [
"tuned_examples/dqn/cartpole-simpleq.yaml",
"tuned_examples/dqn/cartpole-dqn.yaml",
"tuned_examples/dqn/cartpole-dqn-softq.yaml",
"tuned_examples/dqn/cartpole-dqn-param-noise.yaml",
],
args = ["--yaml-dir=tuned_examples/dqn"]
@ -186,6 +187,7 @@ py_test(
data = [
"tuned_examples/dqn/cartpole-simpleq.yaml",
"tuned_examples/dqn/cartpole-dqn.yaml",
"tuned_examples/dqn/cartpole-dqn-softq.yaml",
"tuned_examples/dqn/cartpole-dqn-param-noise.yaml",
],
args = ["--yaml-dir=tuned_examples/dqn", "--torch"]

16
rllib/tuned_examples/dqn/cartpole-dqn-softq.yaml Normal file
View file

@ -0,0 +1,16 @@
cartpole-dqn:
env: CartPole-v0
run: DQN
stop:
episode_reward_mean: 150
timesteps_total: 100000
config:
# Works for both torch and tf.
framework: tf
model:
fcnet_hiddens: [64]
fcnet_activation: linear
n_step: 3
exploration_config:
type: SoftQ
temperature: 0.5