Module neuralops.experiment_utils.model

Module providing model implementations for NeuralOps.

Functions

def load_model(cfg: Any, num_labels: int, device: str = 'auto') ‑> Any

Expand source code

def load_model(cfg: Any, num_labels: int, device: str = "auto") -> Any:
    """Load model using config and num_labels."""
    config = AutoConfig.from_pretrained(cfg.model_name, num_labels=num_labels)
    model = AutoModelForSequenceClassification.from_pretrained(cfg.model_name, config=config)
    fix_layernorm_naming(model)
    if device == "auto":
        device = "cuda" if torch.cuda.is_available() else "cpu"
    model.to(device)
    return model

Load model using config and num_labels.

Classes

class TrainingRun (id: Any = None,
name: str = '',
status: str = 'completed',
created_at: datetime.datetime = <factory>,
config: dict[str, typing.Any] = <factory>,
final_metrics: dict[str, typing.Any] = <factory>,
training_history: dict[str, typing.Any] = <factory>,
actual_duration: float = 0.0,
objects: Any = <neuralops.experiment_utils.model._MockObjectsManager object>)

Expand source code

@dataclass
class TrainingRun:
    """Dataclass representing a training run for comparison and logging."""

    id: Any = None
    name: str = ""
    status: str = "completed"
    created_at: datetime = field(default_factory=datetime.now)
    config: dict[str, Any] = field(default_factory=dict)
    final_metrics: dict[str, Any] = field(default_factory=dict)
    training_history: dict[str, Any] = field(default_factory=dict)
    actual_duration: float = 0.0

    objects: Any = _MockObjectsManager()

Dataclass representing a training run for comparison and logging.

Instance variables

var actual_duration : float

The type of the None singleton.

var config : dict[str, typing.Any]

The type of the None singleton.

var created_at : datetime.datetime

The type of the None singleton.

var final_metrics : dict[str, typing.Any]

The type of the None singleton.

var id : Any

The type of the None singleton.

var name : str

The type of the None singleton.

var objects : Any

The type of the None singleton.

var status : str

The type of the None singleton.

var training_history : dict[str, typing.Any]

The type of the None singleton.