ExperimentConfig#
- class ExperimentConfig(
- device='cuda',
- debug=True,
- deterministic=False,
- seed=0,
- comment='',
- compile=False,
- compile_mode=None,
- compile_compatible_fftconv=False,
- experiment_dir=None,
- num_nodes=1,
- mp_sharing_strategy=None,
- dataset=PLACEHOLDER,
- net=PLACEHOLDER,
- lightning_wrapper_class=PLACEHOLDER,
- optimizer=PLACEHOLDER,
- scheduler=<factory>,
- train=<factory>,
- trainer=<factory>,
- wandb=<factory>,
- start_from_checkpoint=<factory>,
- autoresume=<factory>,
- callbacks=<factory>,
Bases:
objectTop-level configuration for a single nvSubquadratic training run.
All fields have sensible defaults; task-specific overrides are specified in experiment config files under
experiments/. The config is loaded byexperiments/run.py, printed as a tree via Rich, and passed toconstruct_trainer()and the Lightning wrapper.Key optional fields
compile = True: wrap the network withtorch.compile. Mutually exclusive with the QuACK kernel path (useuse_quack=Falsein norm layers when compiling).experiment_dir: override the defaultruns/<run_name>/checkpoint directory with an absolute path.num_nodes: number of multi-node training hosts (passed to the LightningTrainer).
- Parameters:
device (str)
debug (bool)
deterministic (bool)
seed (int)
comment (str)
compile (bool)
compile_mode (str | None)
compile_compatible_fftconv (bool)
experiment_dir (str | None)
num_nodes (int)
mp_sharing_strategy (str | None)
dataset (LazyConfig)
net (LazyConfig)
lightning_wrapper_class (LazyConfig)
optimizer (LazyConfig)
scheduler (SchedulerConfig)
train (TrainConfig)
trainer (TrainerConfig)
wandb (WandbConfig)
start_from_checkpoint (StartFromCheckpointConfig)
autoresume (AutoResumeConfig)
callbacks (list[LazyConfig])
- dataset: LazyConfig = PLACEHOLDER#
- net: LazyConfig = PLACEHOLDER#
- lightning_wrapper_class: LazyConfig = PLACEHOLDER#
- optimizer: LazyConfig = PLACEHOLDER#
- scheduler: SchedulerConfig#
- train: TrainConfig#
- trainer: TrainerConfig#
- wandb: WandbConfig#
- start_from_checkpoint: StartFromCheckpointConfig#
- autoresume: AutoResumeConfig#
- callbacks: list[LazyConfig]#
- __init__(
- device='cuda',
- debug=True,
- deterministic=False,
- seed=0,
- comment='',
- compile=False,
- compile_mode=None,
- compile_compatible_fftconv=False,
- experiment_dir=None,
- num_nodes=1,
- mp_sharing_strategy=None,
- dataset=PLACEHOLDER,
- net=PLACEHOLDER,
- lightning_wrapper_class=PLACEHOLDER,
- optimizer=PLACEHOLDER,
- scheduler=<factory>,
- train=<factory>,
- trainer=<factory>,
- wandb=<factory>,
- start_from_checkpoint=<factory>,
- autoresume=<factory>,
- callbacks=<factory>,
- Parameters:
device (str)
debug (bool)
deterministic (bool)
seed (int)
comment (str)
compile (bool)
compile_mode (str | None)
compile_compatible_fftconv (bool)
experiment_dir (str | None)
num_nodes (int)
mp_sharing_strategy (str | None)
dataset (LazyConfig)
net (LazyConfig)
lightning_wrapper_class (LazyConfig)
optimizer (LazyConfig)
scheduler (SchedulerConfig)
train (TrainConfig)
trainer (TrainerConfig)
wandb (WandbConfig)
start_from_checkpoint (StartFromCheckpointConfig)
autoresume (AutoResumeConfig)
callbacks (list[LazyConfig])
- Return type:
None