mirror of
https://github.com/vale981/ray
synced 2025-03-06 10:31:39 -05:00
cf0894d396
* Halfway done with transferring MAML to new Ray * MAML Beta Out * Debugging MAML atm * Distributed Execution * Pendulum Mass Working * All experiments complete * Cleaned up codebase * Travis CI * Travis CI * Tests * Merged conflicts * Fixed variance bug conflict * Comment resolved * Apply suggestions from code review fixed test_maml * Update rllib/agents/maml/tests/test_maml.py * asdf * Fix testing Co-authored-by: Sven Mika <sven@anyscale.io>
26 lines
665 B
YAML
26 lines
665 B
YAML
pendulum-mass-maml:
|
|
env: ray.rllib.examples.env.pendulum_mass.PendulumMassEnv
|
|
run: MAML
|
|
stop:
|
|
training_iteration: 500
|
|
config:
|
|
horizon: 200
|
|
rollout_fragment_length: 200
|
|
num_envs_per_worker: 10
|
|
inner_adaptation_steps: 1
|
|
maml_optimizer_steps: 5
|
|
gamma: 0.99
|
|
lambda: 1.0
|
|
lr: 0.001
|
|
vf_loss_coeff: 0.5
|
|
clip_param: 0.3
|
|
kl_target: 0.01
|
|
kl_coeff: 0.001
|
|
num_workers: 20
|
|
num_gpus: 1
|
|
inner_lr: 0.03
|
|
explore: True
|
|
clip_actions: False
|
|
model:
|
|
fcnet_hiddens: [64, 64]
|
|
free_log_std: True
|