def from_pretrained(cls,
model_name_or_path,
checkpoint_file="model.pt",
data_name_or_path=".",
bpe="sentencepiece",
**kwargs):
from fairseq_stchde 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,
**kwargs,
)
return RobertaHubInterface(x["args"], x["task"], x["models"][0])
def from_pretrained(
cls,
model_name_or_path,
checkpoint_file="model.pt",
data_name_or_path=".",
**kwargs,
):
"""
Load a :class:`~fairseq_stchde.models.FairseqModel` from a pre-trained model
file. Downloads and caches the pre-trained model file if needed.
The base implementation returns a
:class:`~fairseq_stchde.hub_utils.GeneratorHubInterface`, which can be used to
generate translations or sample from language models. The underlying
:class:`~fairseq_stchde.models.FairseqModel` can be accessed via the
*generator.models* attribute.
Other models may override this to implement custom hub interfaces.
Args:
model_name_or_path (str): either the name of a pre-trained model to
load or a path/URL to a pre-trained model state dict
checkpoint_file (str, optional): colon-separated list of checkpoint
files in the model archive to ensemble (default: 'model.pt')
data_name_or_path (str, optional): point args.data to the archive
at the given path/URL. Can start with '.' or './' to reuse the
model archive path.
"""
from fairseq_stchde import hub_utils
x = hub_utils.from_pretrained(
model_name_or_path,
checkpoint_file,
data_name_or_path,
archive_map=cls.hub_models(),
**kwargs,
)
logger.info(x["args"])
return hub_utils.GeneratorHubInterface(x["args"], x["task"], x["models"])
def from_pretrained(
cls,
model_name_or_path,
checkpoint_file="model.pt",
data_name_or_path=".",
bpe="gpt2",
sample_break_mode="eos",
**kwargs,
):
from fairseq_stchde 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,
sample_break_mode=sample_break_mode,
**kwargs,
)
return BARTHubInterface(x["args"], x["task"], x["models"][0])
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_stchde 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])