Parameter Tuning
This commit is contained in:
@@ -1,8 +1,8 @@
|
||||
_target_: stable_worldmodel.solver.GradientSolver
|
||||
model: ???
|
||||
# Original adam.yaml reference: n_steps=30, num_samples=100, batch_size=1, lr=0.1.
|
||||
n_steps: 125
|
||||
batch_size: 50
|
||||
n_steps: 90
|
||||
batch_size: 100
|
||||
num_samples: 1
|
||||
action_noise: 0
|
||||
device: "cuda"
|
||||
@@ -11,4 +11,4 @@ optimizer_cls:
|
||||
_target_: hydra.utils.get_class
|
||||
path: torch.optim.AdamW
|
||||
optimizer_kwargs:
|
||||
lr: 0.05
|
||||
lr: 0.075
|
||||
|
||||
Reference in New Issue
Block a user