modules/voice_conversion/fairseq/models/roberta/model_gottbert.py

# Copyright (c) Facebook, Inc. and its affiliates.
#
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
GottBERT: a pure German Language Model
"""

from fairseq.models import register_model

from .hub_interface import RobertaHubInterface
from .model import RobertaModel


@register_model("gottbert")
class GottbertModel(RobertaModel):
    @classmethod
    def hub_models(cls):
        return {
            "gottbert-base": "https://dl.gottbert.de/fairseq/models/gottbert-base.tar.gz",
        }

    @classmethod
    def from_pretrained(
        cls,
        model_name_or_path,
        checkpoint_file="model.pt",
        data_name_or_path=".",
        bpe="hf_byte_bpe",
        bpe_vocab="vocab.json",
        bpe_merges="merges.txt",
        bpe_add_prefix_space=False,
        **kwargs
    ):
        from fairseq import hub_utils

        x = hub_utils.from_pretrained(
            model_name_or_path,
            checkpoint_file,
            data_name_or_path,
            archive_map=cls.hub_models(),
            bpe=bpe,
            load_checkpoint_heads=True,
            bpe_vocab=bpe_vocab,
            bpe_merges=bpe_merges,
            bpe_add_prefix_space=bpe_add_prefix_space,
            **kwargs,
        )
        return RobertaHubInterface(x["args"], x["task"], x["models"][0])