pretrained.py

import os
import glob
from typing import Dict, Union, Any

from allennlp.common import Params
from allennlp.predictors import Predictor

from allennlp.common.model_card import ModelCard
from allennlp.common.task_card import TaskCard

# These imports are included so that the model cards can be filled with default information
# obtained from the registered model classes.
from allennlp_models.classification.models import *  # noqa: F401, F403
from allennlp_models.coref.models import *  # noqa: F401, F403
from allennlp_models.generation.models import *  # noqa: F401, F403
from allennlp_models.lm.models import *  # noqa: F401, F403
from allennlp_models.mc.models import *  # noqa: F401, F403
from allennlp_models.pair_classification.models import *  # noqa: F401, F403
from allennlp_models.rc.models import *  # noqa: F401, F403
from allennlp_models.structured_prediction.models import *  # noqa: F401, F403
from allennlp_models.tagging.models import *  # noqa: F401, F403
from allennlp_models.vision.models import *  # noqa: F401, F403


def get_tasks() -> Dict[str, TaskCard]:
    """
    Returns a mapping of [`TaskCard`](/models/common/task_card#taskcard)s for all
    tasks.
    """

    tasks = {}
    task_card_paths = os.path.join(
        os.path.dirname(os.path.realpath(__file__)), "taskcards", "*.json"
    )

    for task_card_path in glob.glob(task_card_paths):
        if "template" not in task_card_path:
            task_card = TaskCard.from_params(params=Params.from_file(task_card_path))
            tasks[task_card.id] = task_card
    return tasks


def get_pretrained_models() -> Dict[str, ModelCard]:
    """
    Returns a mapping of [`ModelCard`](/models/common/model_card#modelcard)s for all
    available pretrained models.
    """

    pretrained_models = {}
    model_card_paths = os.path.join(
        os.path.dirname(os.path.realpath(__file__)), "modelcards", "*.json"
    )

    for model_card_path in glob.glob(model_card_paths):
        if "template" not in model_card_path:
            model_card = ModelCard.from_params(params=Params.from_file(model_card_path))
            pretrained_models[model_card.id] = model_card
    return pretrained_models


def load_predictor(
    model_id: str,
    pretrained_models: Dict[str, ModelCard] = None,
    cuda_device: int = -1,
    overrides: Union[str, Dict[str, Any]] = None,
) -> Predictor:
    """
    Returns the `Predictor` corresponding to the given `model_id`.

    The `model_id` should be key present in the mapping returned by
    [`get_pretrained_models`](#get_pretrained_models).
    """
    pretrained_models = pretrained_models or get_pretrained_models()
    model_card = pretrained_models[model_id]
    if model_card.model_usage.archive_file is None:
        raise ValueError(f"archive_file is required in the {model_card}")
    return Predictor.from_path(
        model_card.model_usage.archive_file,
        predictor_name=model_card.registered_predictor_name,
        cuda_device=cuda_device,
        overrides=overrides,
    )