def from_pretrained(cls, pretrained_model_name_or_path, **kwargs):
r""" Instantiate a :class:`~pytorch_transformers.PretrainedConfig` (or a derived class) from a pre-trained model configuration.
Parameters:
pretrained_model_name_or_path: either:
- a string with the `shortcut name` of a pre-trained model configuration to load from cache or download, e.g.: ``bert-base-uncased``.
- a path to a `directory` containing a configuration file saved using the :func:`~pytorch_transformers.PretrainedConfig.save_pretrained` method, e.g.: ``./my_model_directory/``.
- a path or url to a saved configuration JSON `file`, e.g.: ``./my_model_directory/configuration.json``.
cache_dir: (`optional`) string:
Path to a directory in which a downloaded pre-trained model
configuration should be cached if the standard cache should not be used.
kwargs: (`optional`) dict: key/value pairs with which to update the configuration object after loading.
- The values in kwargs of any keys which are configuration attributes will be used to override the loaded values.
- Behavior concerning key/value pairs whose keys are *not* configuration attributes is controlled by the `return_unused_kwargs` keyword parameter.
force_download: (`optional`) boolean, default False:
Force to (re-)download the model weights and configuration files and override the cached versions if they exists.
proxies: (`optional`) dict, default None:
A dictionary of proxy servers to use by protocol or endpoint, e.g.: {'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}.
The proxies are used on each request.
return_unused_kwargs: (`optional`) bool:
- If False, then this function returns just the final configuration object.
- If True, then this functions returns a tuple `(config, unused_kwargs)` where `unused_kwargs` is a dictionary consisting of the key/value pairs whose keys are not configuration attributes: ie the part of kwargs which has not been used to update `config` and is otherwise ignored.
Examples::
# We can't instantiate directly the base class `PretrainedConfig` so let's show the examples on a
# derived class: BertConfig
config = BertConfig.from_pretrained('bert-base-uncased') # Download configuration from S3 and cache.
config = BertConfig.from_pretrained('./test/saved_model/') # E.g. config (or model) was saved using `save_pretrained('./test/saved_model/')`
config = BertConfig.from_pretrained('./test/saved_model/my_configuration.json')
config = BertConfig.from_pretrained('bert-base-uncased', output_attention=True, foo=False)
assert config.output_attention == True
config, unused_kwargs = BertConfig.from_pretrained('bert-base-uncased', output_attention=True,
foo=False, return_unused_kwargs=True)
assert config.output_attention == True
assert unused_kwargs == {'foo': False}
"""
cache_dir = kwargs.pop('cache_dir', None)
force_download = kwargs.pop('force_download', False)
proxies = kwargs.pop('proxies', None)
return_unused_kwargs = kwargs.pop('return_unused_kwargs', False)
if pretrained_model_name_or_path in cls.pretrained_config_archive_map:
config_file = cls.pretrained_config_archive_map[
pretrained_model_name_or_path]
elif os.path.isdir(pretrained_model_name_or_path):
config_file = os.path.join(pretrained_model_name_or_path,
CONFIG_NAME)
else:
config_file = pretrained_model_name_or_path
# redirect to the cache, if necessary
try:
resolved_config_file = cached_path(config_file,
cache_dir=cache_dir,
force_download=force_download,
proxies=proxies)
except EnvironmentError as e:
if pretrained_model_name_or_path in cls.pretrained_config_archive_map:
logger.error(
"Couldn't reach server at '{}' to download pretrained model configuration file."
.format(config_file))
else:
logger.error(
"Model name '{}' was not found in model name list ({}). "
"We assumed '{}' was a path or url but couldn't find any file "
"associated to this path or url.".format(
pretrained_model_name_or_path,
', '.join(cls.pretrained_config_archive_map.keys()),
config_file))
raise e
if resolved_config_file == config_file:
logger.info("loading configuration file {}".format(config_file))
else:
logger.info(
"loading configuration file {} from cache at {}".format(
config_file, resolved_config_file))
# Load config
config = cls.from_json_file(resolved_config_file)
# Update config with kwargs if needed
to_remove = []
for key, value in kwargs.items():
if hasattr(config, key):
setattr(config, key, value)
to_remove.append(key)
for key in to_remove:
kwargs.pop(key, None)
logger.info("Model config %s", config)
if return_unused_kwargs:
return config, kwargs
else:
return config
def from_pretrained(cls, pretrained_model_name_or_path, *model_args,
**kwargs):
r"""Instantiate a pretrained pytorch model from a pre-trained model configuration.
The model is set in evaluation mode by default using ``model.eval()`` (Dropout modules are deactivated)
To train the model, you should first set it back in training mode with ``model.train()``
The warning ``Weights from XXX not initialized from pretrained model`` means that the weights of XXX do not come pre-trained with the rest of the model.
It is up to you to train those weights with a downstream fine-tuning task.
The warning ``Weights from XXX not used in YYY`` means that the layer XXX is not used by YYY, therefore those weights are discarded.
Parameters:
pretrained_model_name_or_path: either:
- a string with the `shortcut name` of a pre-trained model to load from cache or download, e.g.: ``bert-base-uncased``.
- a path to a `directory` containing model weights saved using :func:`~pytorch_transformers.PreTrainedModel.save_pretrained`, e.g.: ``./my_model_directory/``.
- a path or url to a `tensorflow index checkpoint file` (e.g. `./tf_model/model.ckpt.index`). In this case, ``from_tf`` should be set to True and a configuration object should be provided as ``config`` argument. This loading path is slower than converting the TensorFlow checkpoint in a PyTorch model using the provided conversion scripts and loading the PyTorch model afterwards.
model_args: (`optional`) Sequence of positional arguments:
All remaning positional arguments will be passed to the underlying model's ``__init__`` method
config: (`optional`) instance of a class derived from :class:`~pytorch_transformers.PretrainedConfig`:
Configuration for the model to use instead of an automatically loaded configuation. Configuration can be automatically loaded when:
- the model is a model provided by the library (loaded with the ``shortcut-name`` string of a pretrained model), or
- the model was saved using :func:`~pytorch_transformers.PreTrainedModel.save_pretrained` and is reloaded by suppling the save directory.
- the model is loaded by suppling a local directory as ``pretrained_model_name_or_path`` and a configuration JSON file named `config.json` is found in the directory.
state_dict: (`optional`) dict:
an optional state dictionnary for the model to use instead of a state dictionary loaded from saved weights file.
This option can be used if you want to create a model from a pretrained configuration but load your own weights.
In this case though, you should check if using :func:`~pytorch_transformers.PreTrainedModel.save_pretrained` and :func:`~pytorch_transformers.PreTrainedModel.from_pretrained` is not a simpler option.
cache_dir: (`optional`) string:
Path to a directory in which a downloaded pre-trained model
configuration should be cached if the standard cache should not be used.
force_download: (`optional`) boolean, default False:
Force to (re-)download the model weights and configuration files and override the cached versions if they exists.
proxies: (`optional`) dict, default None:
A dictionary of proxy servers to use by protocol or endpoint, e.g.: {'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}.
The proxies are used on each request.
output_loading_info: (`optional`) boolean:
Set to ``True`` to also return a dictionnary containing missing keys, unexpected keys and error messages.
kwargs: (`optional`) Remaining dictionary of keyword arguments:
Can be used to update the configuration object (after it being loaded) and initiate the model. (e.g. ``output_attention=True``). Behave differently depending on whether a `config` is provided or automatically loaded:
- If a configuration is provided with ``config``, ``**kwargs`` will be directly passed to the underlying model's ``__init__`` method (we assume all relevant updates to the configuration have already been done)
- If a configuration is not provided, ``kwargs`` will be first passed to the configuration class initialization function (:func:`~pytorch_transformers.PretrainedConfig.from_pretrained`). Each key of ``kwargs`` that corresponds to a configuration attribute will be used to override said attribute with the supplied ``kwargs`` value. Remaining keys that do not correspond to any configuration attribute will be passed to the underlying model's ``__init__`` function.
Examples::
model = BertModel.from_pretrained('bert-base-uncased') # Download model and configuration from S3 and cache.
model = BertModel.from_pretrained('./test/saved_model/') # E.g. model was saved using `save_pretrained('./test/saved_model/')`
model = BertModel.from_pretrained('bert-base-uncased', output_attention=True) # Update configuration during loading
assert model.config.output_attention == True
# Loading from a TF checkpoint file instead of a PyTorch model (slower)
config = BertConfig.from_json_file('./tf_model/my_tf_model_config.json')
model = BertModel.from_pretrained('./tf_model/my_tf_checkpoint.ckpt.index', from_tf=True, config=config)
"""
config = kwargs.pop('config', None)
state_dict = kwargs.pop('state_dict', None)
cache_dir = kwargs.pop('cache_dir', None)
from_tf = kwargs.pop('from_tf', False)
force_download = kwargs.pop('force_download', False)
proxies = kwargs.pop('proxies', None)
output_loading_info = kwargs.pop('output_loading_info', False)
# Load config
if config is None:
config, model_kwargs = cls.config_class.from_pretrained(
pretrained_model_name_or_path,
*model_args,
cache_dir=cache_dir,
return_unused_kwargs=True,
force_download=force_download,
**kwargs)
else:
model_kwargs = kwargs
# Load model
if pretrained_model_name_or_path in cls.pretrained_model_archive_map:
archive_file = cls.pretrained_model_archive_map[
pretrained_model_name_or_path]
elif os.path.isdir(pretrained_model_name_or_path):
if from_tf:
# Directly load from a TensorFlow checkpoint
archive_file = os.path.join(pretrained_model_name_or_path,
TF_WEIGHTS_NAME + ".index")
else:
archive_file = os.path.join(pretrained_model_name_or_path,
WEIGHTS_NAME)
else:
if from_tf:
# Directly load from a TensorFlow checkpoint
archive_file = pretrained_model_name_or_path + ".index"
else:
archive_file = pretrained_model_name_or_path
# redirect to the cache, if necessary
try:
resolved_archive_file = cached_path(archive_file,
cache_dir=cache_dir,
force_download=force_download,
proxies=proxies)
except EnvironmentError as e:
if pretrained_model_name_or_path in cls.pretrained_model_archive_map:
logger.error(
"Couldn't reach server at '{}' to download pretrained weights."
.format(archive_file))
else:
logger.error(
"Model name '{}' was not found in model name list ({}). "
"We assumed '{}' was a path or url but couldn't find any file "
"associated to this path or url.".format(
pretrained_model_name_or_path,
', '.join(cls.pretrained_model_archive_map.keys()),
archive_file))
raise e
if resolved_archive_file == archive_file:
logger.info("loading weights file {}".format(archive_file))
else:
logger.info("loading weights file {} from cache at {}".format(
archive_file, resolved_archive_file))
# Instantiate model.
model = cls(config, *model_args, **model_kwargs)
if state_dict is None and not from_tf:
state_dict = torch.load(resolved_archive_file, map_location='cpu')
if from_tf:
# Directly load from a TensorFlow checkpoint
return cls.load_tf_weights(
model, config,
resolved_archive_file[:-6]) # Remove the '.index'
# Convert old format to new format if needed from a PyTorch state_dict
old_keys = []
new_keys = []
for key in state_dict.keys():
new_key = None
if 'gamma' in key:
new_key = key.replace('gamma', 'weight')
if 'beta' in key:
new_key = key.replace('beta', 'bias')
if new_key:
old_keys.append(key)
new_keys.append(new_key)
for old_key, new_key in zip(old_keys, new_keys):
state_dict[new_key] = state_dict.pop(old_key)
# Load from a PyTorch state_dict
missing_keys = []
unexpected_keys = []
error_msgs = []
# copy state_dict so _load_from_state_dict can modify it
metadata = getattr(state_dict, '_metadata', None)
state_dict = state_dict.copy()
if metadata is not None:
state_dict._metadata = metadata
def load(module, prefix=''):
local_metadata = {} if metadata is None else metadata.get(
prefix[:-1], {})
module._load_from_state_dict(state_dict, prefix, local_metadata,
True, missing_keys, unexpected_keys,
error_msgs)
for name, child in module._modules.items():
if child is not None:
load(child, prefix + name + '.')
# Make sure we are able to load base models as well as derived models (with heads)
start_prefix = ''
model_to_load = model
if not hasattr(model, cls.base_model_prefix) and any(
s.startswith(cls.base_model_prefix)
for s in state_dict.keys()):
start_prefix = cls.base_model_prefix + '.'
if hasattr(model, cls.base_model_prefix) and not any(
s.startswith(cls.base_model_prefix)
for s in state_dict.keys()):
model_to_load = getattr(model, cls.base_model_prefix)
load(model_to_load, prefix=start_prefix)
if len(missing_keys) > 0:
logger.info(
"Weights of {} not initialized from pretrained model: {}".
format(model.__class__.__name__, missing_keys))
if len(unexpected_keys) > 0:
logger.info(
"Weights from pretrained model not used in {}: {}".format(
model.__class__.__name__, unexpected_keys))
if len(error_msgs) > 0:
raise RuntimeError(
'Error(s) in loading state_dict for {}:\n\t{}'.format(
model.__class__.__name__, "\n\t".join(error_msgs)))
if hasattr(model, 'tie_weights'):
model.tie_weights(
) # make sure word embedding weights are still tied
# Set model in evaluation mode to desactivate DropOut modules by default
model.eval()
if output_loading_info:
loading_info = {
"missing_keys": missing_keys,
"unexpected_keys": unexpected_keys,
"error_msgs": error_msgs
}
return model, loading_info
return model
def from_pretrained(cls, pretrained_model_name_or_path, *inputs, **kwargs):
"""
Instantiate a BertPreTrainedModel from a pre-trained model file or a pytorch state dict.
Download and cache the pre-trained model file if needed.
Params:
pretrained_model_name_or_path: either:
- a str with the name of a pre-trained model to load selected in the list of:
. `bert-base-uncased`
. `bert-large-uncased`
. `bert-base-cased`
. `bert-large-cased`
. `bert-base-multilingual-uncased`
. `bert-base-multilingual-cased`
. `bert-base-chinese`
- a path or url to a pretrained model archive containing:
. `bert_config.json` a configuration file for the model
. `pytorch_model.bin` a PyTorch dump of a BertForPreTraining instance
- a path or url to a pretrained model archive containing:
. `bert_config.json` a configuration file for the model
. `model.chkpt` a TensorFlow checkpoint
from_tf: should we load the weights from a locally saved TensorFlow checkpoint
cache_dir: an optional path to a folder in which the pre-trained models will be cached.
state_dict: an optional state dictionnary (collections.OrderedDict object) to use instead of Google pre-trained models
*inputs, **kwargs: additional input for the specific Bert class
(ex: num_labels for BertForSequenceClassification)
"""
state_dict = kwargs.get('state_dict', None)
kwargs.pop('state_dict', None)
cache_dir = kwargs.get('cache_dir', None)
kwargs.pop('cache_dir', None)
from_tf = kwargs.get('from_tf', False)
kwargs.pop('from_tf', None)
if pretrained_model_name_or_path in PRETRAINED_MODEL_ARCHIVE_MAP:
archive_file = PRETRAINED_MODEL_ARCHIVE_MAP[pretrained_model_name_or_path]
else:
archive_file = pretrained_model_name_or_path
# redirect to the cache, if necessary
try:
resolved_archive_file = cached_path(archive_file, cache_dir=cache_dir)
except EnvironmentError:
logger.error(
"Model name '{}' was not found in model name list ({}). "
"We assumed '{}' was a path or url but couldn't find any file "
"associated to this path or url.".format(
pretrained_model_name_or_path,
', '.join(PRETRAINED_MODEL_ARCHIVE_MAP.keys()),
archive_file))
return None
if resolved_archive_file == archive_file:
logger.info("loading archive file {}".format(archive_file))
else:
logger.info("loading archive file {} from cache at {}".format(
archive_file, resolved_archive_file))
tempdir = None
if os.path.isdir(resolved_archive_file) or from_tf:
serialization_dir = resolved_archive_file
else:
# Extract archive to temp dir
tempdir = tempfile.mkdtemp()
logger.info("extracting archive file {} to temp dir {}".format(
resolved_archive_file, tempdir))
with tarfile.open(resolved_archive_file, 'r:gz') as archive:
archive.extractall(tempdir)
serialization_dir = tempdir
# Load config
config_file = os.path.join(serialization_dir, CONFIG_NAME)
if not os.path.exists(config_file):
# Backward compatibility with old naming format
config_file = os.path.join(serialization_dir, BERT_CONFIG_NAME)
config = BertConfig.from_json_file(config_file)
logger.info("Model config {}".format(config))
# Instantiate model.
model = cls(config, *inputs, **kwargs)
if state_dict is None and not from_tf:
weights_path = os.path.join(serialization_dir, WEIGHTS_NAME)
state_dict = torch.load(weights_path, map_location='cpu')
if tempdir:
# Clean up temp dir
shutil.rmtree(tempdir)
if from_tf:
# Directly load from a TensorFlow checkpoint
weights_path = os.path.join(serialization_dir, TF_WEIGHTS_NAME)
return load_tf_weights_in_bert(model, weights_path)
# Load from a PyTorch state_dict
old_keys = []
new_keys = []
for key in state_dict.keys():
new_key = None
if 'gamma' in key:
new_key = key.replace('gamma', 'weight')
if 'beta' in key:
new_key = key.replace('beta', 'bias')
if new_key:
old_keys.append(key)
new_keys.append(new_key)
for old_key, new_key in zip(old_keys, new_keys):
state_dict[new_key] = state_dict.pop(old_key)
missing_keys = []
unexpected_keys = []
error_msgs = []
# copy state_dict so _load_from_state_dict can modify it
metadata = getattr(state_dict, '_metadata', None)
state_dict = state_dict.copy()
if metadata is not None:
state_dict._metadata = metadata
def load(module, prefix=''):
local_metadata = {} if metadata is None else metadata.get(prefix[:-1], {})
module._load_from_state_dict(
state_dict, prefix, local_metadata, False, missing_keys, unexpected_keys, error_msgs)
for name, child in module._modules.items():
if child is not None:
load(child, prefix + name + '.')
start_prefix = ''
if not hasattr(model, 'bert') and any(s.startswith('bert.') for s in state_dict.keys()):
start_prefix = 'bert.'
load(model, prefix=start_prefix)
if len(missing_keys) > 0:
logger.info("Weights of {} not initialized from pretrained model: {}".format(
model.__class__.__name__, missing_keys))
if len(unexpected_keys) > 0:
logger.info("Weights from pretrained model not used in {}: {}".format(
model.__class__.__name__, unexpected_keys))
if len(error_msgs) > 0:
raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format(
model.__class__.__name__, "\n\t".join(error_msgs)))
return model
from pytorch_pretrained_bert.file_utils import cached_path
import os
import shutil
files = [
'https://raw.githubusercontent.com/glample/tagger/master/dataset/eng.testa',
'https://raw.githubusercontent.com/glample/tagger/master/dataset/eng.testb',
'https://raw.githubusercontent.com/glample/tagger/master/dataset/eng.train',
]
target_dir = 'data/conll2003'
os.makedirs(target_dir, exist_ok=True)
for fname in files:
name = os.path.basename(fname)
shutil.copyfile(cached_path(fname), os.path.join(target_dir, name))
print(fname, '=>', name)
def from_pretrained(cls,
pretrained_model_name,
cache_dir=None,
*inputs,
**kwargs):
"""
Instantiate a PreTrainedBertModel from a pre-trained model file.
Download and cache the pre-trained model file if needed.
Params:
pretrained_model_name: either:
- a str with the name of a pre-trained model to load selected in the list of:
. `bert-base-uncased`
. `bert-large-uncased`
. `bert-base-cased`
. `bert-base-multilingual`
. `bert-base-chinese`
- a path or url to a pretrained model archive containing:
. `bert_config.json` a configuration file for the model
. `pytorch_model.bin` a PyTorch dump of a BertForPreTraining instance
*inputs, **kwargs: additional input for the specific Bert class
(ex: num_labels for BertForSequenceClassification)
"""
if pretrained_model_name in PRETRAINED_MODEL_ARCHIVE_MAP:
archive_file = PRETRAINED_MODEL_ARCHIVE_MAP[pretrained_model_name]
else:
archive_file = pretrained_model_name
# redirect to the cache, if necessary
try:
resolved_archive_file = cached_path(archive_file,
cache_dir=cache_dir)
except FileNotFoundError:
logger.error(
"Model name '{}' was not found in model name list ({}). "
"We assumed '{}' was a path or url but couldn't find any file "
"associated to this path or url.".format(
pretrained_model_name,
', '.join(PRETRAINED_MODEL_ARCHIVE_MAP.keys()),
pretrained_model_name))
return None
if resolved_archive_file == archive_file:
logger.info("loading archive file {}".format(archive_file))
else:
logger.info("loading archive file {} from cache at {}".format(
archive_file, resolved_archive_file))
tempdir = None
if os.path.isdir(resolved_archive_file):
serialization_dir = resolved_archive_file
else:
# Extract archive to temp dir
tempdir = tempfile.mkdtemp()
logger.info("extracting archive file {} to temp dir {}".format(
resolved_archive_file, tempdir))
with tarfile.open(resolved_archive_file, 'r:gz') as archive:
archive.extractall(tempdir)
serialization_dir = tempdir
# Load config
config_file = os.path.join(serialization_dir, CONFIG_NAME)
config = BertConfig.from_json_file(config_file)
logger.info("Model config {}".format(config))
# Instantiate model.
model = cls(config, *inputs, **kwargs)
weights_path = os.path.join(serialization_dir, WEIGHTS_NAME)
state_dict = torch.load(weights_path)
missing_keys = []
unexpected_keys = []
error_msgs = []
# copy state_dict so _load_from_state_dict can modify it
metadata = getattr(state_dict, '_metadata', None)
state_dict = state_dict.copy()
if metadata is not None:
state_dict._metadata = metadata
def load(module, prefix=''):
local_metadata = {} if metadata is None else metadata.get(
prefix[:-1], {})
module._load_from_state_dict(state_dict, prefix, local_metadata,
True, missing_keys, unexpected_keys,
error_msgs)
for name, child in module._modules.items():
if child is not None:
load(child, prefix + name + '.')
load(model, prefix='' if hasattr(model, 'bert') else 'bert.')
if len(missing_keys) > 0:
logger.info(
"Weights of {} not initialized from pretrained model: {}".
format(model.__class__.__name__, missing_keys))
if len(unexpected_keys) > 0:
logger.info(
"Weights from pretrained model not used in {}: {}".format(
model.__class__.__name__, unexpected_keys))
if tempdir:
# Clean up temp dir
shutil.rmtree(tempdir)
return model
def from_pretrained(
cls, pretrained_model_name_or_path, num_special_tokens=None, state_dict=None, cache_dir=None, from_tf=False, *inputs, **kwargs
):
"""
Instantiate a OpenAIGPTPreTrainedModel from a pre-trained model file or a pytorch state dict.
Download and cache the pre-trained model file if needed.
Params:
pretrained_model_name_or_path: either:
- a str with the name of a pre-trained model to load selected in the list of:
. `openai-gpt`
- a path or url to a pretrained model archive containing:
. `openai_gpt_config.json` a configuration file for the model
. `pytorch_model.bin` a PyTorch dump of a OpenAIGPTModel instance
- a path or url to a pretrained model archive containing:
. `bert_config.json` a configuration file for the model
. a series of NumPy files containing OpenAI TensorFlow trained weights
from_tf: should we load the weights from a locally saved TensorFlow checkpoint
cache_dir: an optional path to a folder in which the pre-trained models will be cached.
state_dict: an optional state dictionnary (collections.OrderedDict object) to use instead of pre-trained models
*inputs, **kwargs: additional input for the specific Bert class
(ex: num_labels for BertForSequenceClassification)
"""
if pretrained_model_name_or_path in PRETRAINED_MODEL_ARCHIVE_MAP:
archive_file = PRETRAINED_MODEL_ARCHIVE_MAP[pretrained_model_name_or_path]
config_file = PRETRAINED_CONFIG_ARCHIVE_MAP[pretrained_model_name_or_path]
else:
archive_file = os.path.join(pretrained_model_name_or_path, WEIGHTS_NAME)
config_file = os.path.join(pretrained_model_name_or_path, CONFIG_NAME)
# redirect to the cache, if necessary
try:
resolved_archive_file = cached_path(archive_file, cache_dir=cache_dir)
resolved_config_file = cached_path(config_file, cache_dir=cache_dir)
except EnvironmentError:
logger.error(
"Model name '{}' was not found in model name list ({}). "
"We assumed '{}' was a path or url but couldn't find files {} and {} "
"at this path or url.".format(
pretrained_model_name_or_path, ", ".join(PRETRAINED_MODEL_ARCHIVE_MAP.keys()), pretrained_model_name_or_path,
archive_file, config_file
)
)
return None
if resolved_archive_file == archive_file and resolved_config_file == config_file:
logger.info("loading weights file {}".format(archive_file))
logger.info("loading configuration file {}".format(config_file))
else:
logger.info("loading weights file {} from cache at {}".format(
archive_file, resolved_archive_file))
logger.info("loading configuration file {} from cache at {}".format(
config_file, resolved_config_file))
# Load config
config = OpenAIGPTConfig.from_json_file(resolved_config_file)
logger.info("Model config {}".format(config))
# Instantiate model.
model = cls(config, *inputs, **kwargs)
if state_dict is None and not from_tf:
state_dict = torch.load(resolved_archive_file, map_location='cpu' if not torch.cuda.is_available() else None)
if from_tf:
# Directly load from a TensorFlow checkpoint (stored as NumPy array)
return load_tf_weights_in_openai_gpt(model, resolved_archive_file)
old_keys = []
new_keys = []
for key in state_dict.keys():
new_key = None
if key.endswith(".g"):
new_key = key[:-2] + ".weight"
elif key.endswith(".b"):
new_key = key[:-2] + ".bias"
elif key.endswith(".w"):
new_key = key[:-2] + ".weight"
if new_key:
old_keys.append(key)
new_keys.append(new_key)
for old_key, new_key in zip(old_keys, new_keys):
state_dict[new_key] = state_dict.pop(old_key)
missing_keys = []
unexpected_keys = []
error_msgs = []
# copy state_dict so _load_from_state_dict can modify it
metadata = getattr(state_dict, "_metadata", None)
state_dict = state_dict.copy()
if metadata is not None:
state_dict._metadata = metadata
def load(module, prefix=""):
local_metadata = {} if metadata is None else metadata.get(prefix[:-1], {})
module._load_from_state_dict(
state_dict, prefix, local_metadata, True, missing_keys, unexpected_keys, error_msgs
)
for name, child in module._modules.items():
if child is not None:
load(child, prefix + name + ".")
start_model = model
if hasattr(model, "transformer") and all(not s.startswith('transformer.') for s in state_dict.keys()):
start_model = model.transformer
load(start_model, prefix="")
if len(missing_keys) > 0:
logger.info(
"Weights of {} not initialized from pretrained model: {}".format(model.__class__.__name__, missing_keys)
)
if len(unexpected_keys) > 0:
logger.info(
"Weights from pretrained model not used in {}: {}".format(model.__class__.__name__, unexpected_keys)
)
if len(error_msgs) > 0:
raise RuntimeError(
"Error(s) in loading state_dict for {}:\n\t{}".format(model.__class__.__name__, "\n\t".join(error_msgs))
)
# Add additional embeddings for special tokens if needed
# This step also make sure we are still sharing the output and input embeddings after loading weights
model.set_num_special_tokens(num_special_tokens if num_special_tokens is not None else config.n_special)
return model
def from_pretrained(cls,
pretrained_model_name_or_path,
cache_dir=None,
*inputs,
**kwargs):
"""
Instantiate a PreTrainedBertModel from a pre-trained model file.
Download and cache the pre-trained model file if needed.
"""
if pretrained_model_name_or_path in PRETRAINED_VOCAB_ARCHIVE_MAP:
vocab_file = PRETRAINED_VOCAB_ARCHIVE_MAP[
pretrained_model_name_or_path]
merges_file = PRETRAINED_MERGES_ARCHIVE_MAP[
pretrained_model_name_or_path]
special_tokens_file = None
else:
vocab_file = os.path.join(pretrained_model_name_or_path,
VOCAB_NAME)
merges_file = os.path.join(pretrained_model_name_or_path,
MERGES_NAME)
special_tokens_file = os.path.join(pretrained_model_name_or_path,
SPECIAL_TOKENS_NAME)
if not os.path.exists(special_tokens_file):
special_tokens_file = None
else:
logger.info("loading special tokens file {}".format(
special_tokens_file))
# redirect to the cache, if necessary
try:
resolved_vocab_file = cached_path(vocab_file, cache_dir=cache_dir)
resolved_merges_file = cached_path(merges_file,
cache_dir=cache_dir)
except EnvironmentError:
if pretrained_model_name_or_path in PRETRAINED_VOCAB_ARCHIVE_MAP:
logger.error(
"Couldn't reach server at '{}' to download vocabulary.".
format(vocab_file))
else:
logger.error(
"Model name '{}' was not found in model name list ({}). "
"We assumed '{}' was a path or url but couldn't find files {} and {} "
"at this path or url.".format(
pretrained_model_name_or_path,
', '.join(PRETRAINED_VOCAB_ARCHIVE_MAP.keys()),
pretrained_model_name_or_path, vocab_file,
merges_file))
return None
if resolved_vocab_file == vocab_file and resolved_merges_file == merges_file:
logger.info("loading vocabulary file {}".format(vocab_file))
logger.info("loading merges file {}".format(merges_file))
else:
logger.info("loading vocabulary file {} from cache at {}".format(
vocab_file, resolved_vocab_file))
logger.info("loading merges file {} from cache at {}".format(
merges_file, resolved_merges_file))
if pretrained_model_name_or_path in PRETRAINED_VOCAB_POSITIONAL_EMBEDDINGS_SIZE_MAP:
# if we're using a pretrained model, ensure the tokenizer wont index sequences longer
# than the number of positional embeddings
max_len = PRETRAINED_VOCAB_POSITIONAL_EMBEDDINGS_SIZE_MAP[
pretrained_model_name_or_path]
kwargs['max_len'] = min(kwargs.get('max_len', int(1e12)), max_len)
# Instantiate tokenizer.
if special_tokens_file and 'special_tokens' not in kwargs:
special_tokens = open(special_tokens_file,
encoding='utf-8').read().split('\n')[:-1]
else:
special_tokens = kwargs.pop('special_tokens', [])
tokenizer = cls(resolved_vocab_file,
resolved_merges_file,
special_tokens=special_tokens,
*inputs,
**kwargs)
return tokenizer