dair_pll.experiment_config
Configuration dataclasses for experiments.
- class dair_pll.experiment_config.SystemConfig[source]
Bases:
object
Dummy base
dataclass
for parameters for learning dynamics; all inheriting classes are expected to contain all necessary configuration attributes.
- class dair_pll.experiment_config.OptimizerConfig(optimizer=torch.optim.Adam, lr=Float(1e-05), wd=Float(4e-05), epochs=10000, patience=30, batch_size=Int(64))[source]
Bases:
object
dataclass()
defining setup and usage opf a PytorchOptimizer()
for learning.-
lr:
dair_pll.hyperparameter.Float
= Float(1e-05) Learning rate.
-
wd:
dair_pll.hyperparameter.Float
= Float(4e-05) Weight decay.
-
batch_size:
dair_pll.hyperparameter.Int
= Int(64) Size of batch for an individual gradient step.
-
lr:
- class dair_pll.experiment_config.SupervisedLearningExperimentConfig(data_config=<factory>, base_config=<factory>, learnable_config=<factory>, optimizer_config=<factory>, storage='./', run_name='experiment_run', run_wandb=True, wandb_project=None, full_evaluation_period=1, full_evaluation_samples=5, update_geometry_in_videos=False)[source]
Bases:
object
dataclass
defining setup of aSupervisedLearningExperiment
-
data_config:
dair_pll.data_config.DataConfig
Configuration for experiment’s
SystemDataManager
.
-
base_config:
dair_pll.experiment_config.SystemConfig
Configuration for experiment’s “base” system, from which trajectories are modeled and optionally generated.
-
learnable_config:
dair_pll.experiment_config.SystemConfig
Configuration for system to be learned.
-
optimizer_config:
dair_pll.experiment_config.OptimizerConfig
Configuration for experiment’s optimization process.
-
wandb_project:
typing.Optional
[str
] = None If
run_wandb
, a project to store results under on W&B.
-
data_config: