maulberto3/run_mlm_big_text_files.py

## run_mlm_big_text_files.py
"""
This is a workaround for `examples/run_mlm.py` for pretraining models
with big text files line-by-line.

For the time being, `datasets` is facing some issues dealing with really
big text files, so we use a custom dataset until this is fixed.

August 3th 2021

Author: Juan Manuel Pérez

"""


#!/usr/bin/env python
# coding=utf-8
# Copyright 2020 The HuggingFace Team All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
Fine-tuning the library models for masked language modeling (BERT, ALBERT, RoBERTa...) on a text file or a dataset.

Here is the full list of checkpoints on the hub that can be fine-tuned by this script:
https://huggingface.co/models?filter=masked-lm
"""
# You can also adapt this script on your own masked language modeling task. Pointers for this are left as comments.

import logging
import math
from multiprocessing.sharedctypes import Value
import os
import sys
from dataclasses import dataclass, field
from typing import Optional
from glob import glob
import datasets
from datasets import load_dataset
import random
import transformers
from transformers import (
    CONFIG_MAPPING,
    MODEL_FOR_MASKED_LM_MAPPING,
    AutoConfig,
    AutoModelForMaskedLM,
    AutoTokenizer,
    DataCollatorForLanguageModeling,
    HfArgumentParser,
    Trainer,
    TrainingArguments,
    set_seed,
)
from transformers.trainer_utils import get_last_checkpoint
from transformers.utils import check_min_version
from transformers.utils.versions import require_version
from torch.utils.data import IterableDataset, Dataset

# Will error if the minimal version of Transformers is not installed. Remove at your own risks.
check_min_version("4.9.0")

require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/language-modeling/requirements.txt")

logger = logging.getLogger(__name__)
MODEL_CONFIG_CLASSES = list(MODEL_FOR_MASKED_LM_MAPPING.keys())
MODEL_TYPES = tuple(conf.model_type for conf in MODEL_CONFIG_CLASSES)


class BatchProcessedDataset(IterableDataset):
    """
    A dataset which streams and process tweets from files
    """
    def __init__(self, files, tokenizer, batch_size=4096, limit=-1, padding='max_length'):
        """
        Constructor

        Arguments:
        ----------


        files: list[str]
            A list of files from which lines are streamed in order

        tokenizer:
            A transformer's tokenizer

        batch_size:
            The batch size of lines that is passed to the tokenizer

        limit: int (optional)
            An optional limit to the number of streamed lines. By default no limit is set

        padding:

        """
        self.files = files
        self.batch_size = batch_size
        self.tokenizer = tokenizer
        self.padding = padding
        self.limit = limit

    def __nextbatch(self, f):
        nb = [x.strip("\n") for _, x in zip(range(self.batch_size), f)]
        return nb

    def __iter__(self):
        num_iter = 0
        for file_path in self.files:
            logger.info(f"Opening file {file_path}")
            with open(file_path) as f:
                next_batch = self.__nextbatch(f)
                while next_batch:
                    tokenized_batch = self.tokenizer(next_batch, padding=self.padding, truncation=True, return_special_tokens_mask=True)
                    for encoding in tokenized_batch.encodings:
                        if num_iter > 0 and num_iter >= self.limit:
                            return
                        yield {
                            "input_ids": encoding.ids,
                            "token_type_ids": encoding.type_ids,
                            "attention_mask": encoding.attention_mask,
                            "special_tokens_mask": encoding.special_tokens_mask
                        }
                        num_iter += 1
                    next_batch = self.__nextbatch(f)


@dataclass
class ModelArguments:
    """
    Arguments pertaining to which model/config/tokenizer we are going to fine-tune, or train from scratch.
    """

    model_name_or_path: Optional[str] = field(
        default=None,
        metadata={
            "help": "The model checkpoint for weights initialization."
            "Don't set if you want to train a model from scratch."
        },
    )
    model_type: Optional[str] = field(
        default=None,
        metadata={"help": "If training from scratch, pass a model type from the list: " + ", ".join(MODEL_TYPES)},
    )
    config_overrides: Optional[str] = field(
        default=None,
        metadata={
            "help": "Override some existing default config settings when a model is trained from scratch. Example: "
            "n_embd=10,resid_pdrop=0.2,scale_attn_weights=false,summary_type=cls_index"
        },
    )
    config_name: Optional[str] = field(
        default=None, metadata={"help": "Pretrained config name or path if not the same as model_name"}
    )
    tokenizer_name: Optional[str] = field(
        default=None, metadata={"help": "Pretrained tokenizer name or path if not the same as model_name"}
    )


    cache_dir: Optional[str] = field(
        default=None,
        metadata={"help": "Where do you want to store the pretrained models downloaded from huggingface.co"},
    )
    use_fast_tokenizer: bool = field(
        default=True,
        metadata={"help": "Whether to use one of the fast tokenizer (backed by the tokenizers library) or not."},
    )
    model_revision: str = field(
        default="main",
        metadata={"help": "The specific model version to use (can be a branch name, tag name or commit id)."},
    )
    use_auth_token: bool = field(
        default=False,
        metadata={
            "help": "Will use the token generated when running `transformers-cli login` (necessary to use this script "
            "with private models)."
        },
    )


    def __post_init__(self):
        if self.config_overrides is not None and (self.config_name is not None or self.model_name_or_path is not None):
            raise ValueError(
                "--config_overrides can't be used in combination with --config_name or --model_name_or_path"
            )


@dataclass
class DataTrainingArguments:
    """
    Arguments pertaining to what data we are going to input our model for training and eval.
    """

    train_dir: Optional[str] = field(
        default=None,
        metadata={"help": "Path to train files."}
    )
    eval_dir: Optional[str] = field(
        default=None,
        metadata={"help": "Path to test files."}
    )

    max_seq_length: Optional[int] = field(
        default=None,
        metadata={
            "help": "The maximum total input sequence length after tokenization. Sequences longer "
            "than this will be truncated."
        },
    )
    preprocessing_num_workers: Optional[int] = field(
        default=None,
        metadata={"help": "The number of processes to use for the preprocessing."},
    )
    mlm_probability: float = field(
        default=0.15, metadata={"help": "Ratio of tokens to mask for masked language modeling loss"}
    )

    pad_to_max_length: bool = field(
        default=False,
        metadata={
            "help": "Whether to pad all samples to `max_seq_length`. "
            "If False, will pad the samples dynamically when batching to the maximum length in the batch."
        },
    )
    max_train_samples: Optional[int] = field(
        default=None,
        metadata={
            "help": "For debugging purposes or quicker training, truncate the number of training examples to this "
            "value if set."
        },
    )
    max_eval_samples: Optional[int] = field(
        default=None,
        metadata={
            "help": "For debugging purposes or quicker training, truncate the number of evaluation examples to this "
            "value if set."
        },
    )

    tokenization_batch_size: Optional[int] = field(
        default=4096, metadata={
            "help": "Tokenization batch size (for on-the-fly tokenization)"
        }
    )

    tokenize_on_the_fly: bool = field(
        default=False, metadata={
            "help": "Tokenize on-the-fly"
        }
    )

    def __post_init__(self):
        if (self.train_dir is None or self.eval_dir is None:
            raise ValueError("Must provide train_dir & eval_dir")


def main():
    # See all possible arguments in src/transformers/training_args.py
    # or by passing the --help flag to this script.
    # We now keep distinct sets of args, for a cleaner separation of concerns.

    parser = HfArgumentParser((ModelArguments, DataTrainingArguments, TrainingArguments))
    if (2 <= len(sys.argv) <= 4):
        if sys.argv[-1].endswith(".json"):
            json_file = sys.argv[-1]
        else:
            json_file = sys.argv[1]
        # If we pass only one argument to the script and it's the path to a json file,
        # let's parse it to get our arguments.
        model_args, data_args, training_args = parser.parse_json_file(json_file=os.path.abspath(json_file))
    else:
        model_args, data_args, training_args = parser.parse_args_into_dataclasses()

    print("Model args: ",model_args)
    # Setup logging
    logging.basicConfig(
        format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",
        datefmt="%m/%d/%Y %H:%M:%S",
        handlers=[logging.StreamHandler(sys.stdout)],
    )

    log_level = training_args.get_process_log_level()
    logger.setLevel(log_level)
    datasets.utils.logging.set_verbosity(log_level)
    transformers.utils.logging.set_verbosity(log_level)
    transformers.utils.logging.enable_default_handler()
    transformers.utils.logging.enable_explicit_format()

    # Log on each process the small summary:
    logger.warning(
        f"Process rank: {training_args.local_rank}, device: {training_args.device}, n_gpu: {training_args.n_gpu}"
        + f"distributed training: {bool(training_args.local_rank != -1)}, 16-bits training: {training_args.fp16}"
    )
    # Set the verbosity to info of the Transformers logger (on main process only):
    logger.info(f"Training/evaluation parameters {training_args}")

    # Detecting last checkpoint.
    last_checkpoint = None
    if os.path.isdir(training_args.output_dir) and training_args.do_train and not training_args.overwrite_output_dir:
        last_checkpoint = get_last_checkpoint(training_args.output_dir)
        if last_checkpoint is None and len(os.listdir(training_args.output_dir)) > 0:
            raise ValueError(
                f"Output directory ({training_args.output_dir}) already exists and is not empty. "
                "Use --overwrite_output_dir to overcome."
            )
        elif last_checkpoint is not None and training_args.resume_from_checkpoint is None:
            logger.info(
                f"Checkpoint detected, resuming training at {last_checkpoint}. To avoid this behavior, change "
                "the `--output_dir` or add `--overwrite_output_dir` to train from scratch."
            )

    # Set seed before initializing model.
    set_seed(training_args.seed)

    config_kwargs = {
        "cache_dir": model_args.cache_dir,
        "revision": model_args.model_revision,
        "use_auth_token": True if model_args.use_auth_token else None,
    }
    if model_args.config_name:
        config = AutoConfig.from_pretrained(model_args.config_name, **config_kwargs)
    elif model_args.model_name_or_path:
        config = AutoConfig.from_pretrained(model_args.model_name_or_path, **config_kwargs)
    else:
        config = CONFIG_MAPPING[model_args.model_type]()
        logger.warning("You are instantiating a new config instance from scratch.")
        if model_args.config_overrides is not None:
            logger.info(f"Overriding config: {model_args.config_overrides}")
            config.update_from_string(model_args.config_overrides)

    tokenizer_kwargs = {
        "cache_dir": model_args.cache_dir,
        "use_fast": model_args.use_fast_tokenizer,
        "revision": model_args.model_revision,
        "use_auth_token": True if model_args.use_auth_token else None,
    }
    if model_args.tokenizer_name:
        tokenizer = AutoTokenizer.from_pretrained(model_args.tokenizer_name, **tokenizer_kwargs)
    elif model_args.model_name_or_path:
        tokenizer = AutoTokenizer.from_pretrained(model_args.model_name_or_path, **tokenizer_kwargs)
    else:
        raise ValueError(
            "You are instantiating a new tokenizer from scratch. This is not supported by this script."
            "You can do it from another script, save it, and load it from here, using --tokenizer_name."
        )

    if model_args.model_name_or_path:
        model = AutoModelForMaskedLM.from_pretrained(
            model_args.model_name_or_path,
            from_tf=bool(".ckpt" in model_args.model_name_or_path),
            config=config,
            cache_dir=model_args.cache_dir,
            revision=model_args.model_revision,
            use_auth_token=True if model_args.use_auth_token else None,
        )
    else:
        logger.info("Training new model from scratch")
        model = AutoModelForMaskedLM.from_config(config)

    model.resize_token_embeddings(len(tokenizer))
    tokenizer.model_max_length =128
    # Get the datasets: you can either provide your own CSV/JSON/TXT training and evaluation files (see below)
    # or just provide the name of one of the public datasets available on the hub at https://huggingface.co/datasets/
    # (the dataset will be downloaded automatically from the datasets Hub
    #
    # For CSV/JSON files, this script will use the column called 'text' or the first column. You can easily tweak this
    # behavior (see below)
    #
    # In distributed training, the load_dataset function guarantee that only one local process can concurrently
    # download the dataset.
    padding = "max_length" if data_args.pad_to_max_length else False

    train_files = glob(os.path.join(data_args.train_dir, "*.txt"))
    eval_files = glob(os.path.join(data_args.eval_dir, "*.txt"))

    if data_args.tokenize_on_the_fly:
        """
        Use our custom class
        """
        random.shuffle(train_files)

        if data_args.max_eval_samples is None or not type(data_args.max_eval_samples) is int:
            raise ValueError("Must provide max_eval_samples")
        logger.info(f"Tokenization batch size {data_args.tokenization_batch_size}")
        train_dataset = BatchProcessedDataset(
            train_files, tokenizer, batch_size=data_args.tokenization_batch_size,
            padding=padding,
        )
        eval_dataset = BatchProcessedDataset(
            eval_files, tokenizer, batch_size=data_args.tokenization_batch_size,
            padding=padding, limit=data_args.max_eval_samples
        )

    else:
        """
        Load using datasets
        """

        raw_datasets = load_dataset("text",
            data_files = {"train": train_files, "test": eval_files}
        )


    if data_args.max_seq_length is None:
        max_seq_length = tokenizer.model_max_length
        if max_seq_length > 1024:
            logger.warning(
                f"The tokenizer picked seems to have a very large `model_max_length` ({tokenizer.model_max_length}). "
                "Picking 1024 instead. You can change that default value by passing --max_seq_length xxx."
            )
            max_seq_length = 1024
    else:
        if data_args.max_seq_length > tokenizer.model_max_length:
            logger.warning(
                f"The max_seq_length passed ({data_args.max_seq_length}) is larger than the maximum length for the"
                f"model ({tokenizer.model_max_length}). Using max_seq_length={tokenizer.model_max_length}."
            )
        max_seq_length = min(data_args.max_seq_length, tokenizer.model_max_length)

    if data_args.tokenize_on_the_fly:
        training_args.remove_unused_columns = False
    else:
        def tokenize_function(examples):
            # Remove empty lines
            examples["text"] = [
                line for line in examples["text"] if len(line) > 0 and not line.isspace()
            ]
            return tokenizer(
                examples["text"],
                padding=padding,
                truncation=True,
                max_length=max_seq_length,
                # We use this option because DataCollatorForLanguageModeling (see below) is more efficient when it
                # receives the `special_tokens_mask`.
                return_special_tokens_mask=True,
            )

        with training_args.main_process_first(desc="dataset map tokenization"):
            tokenized_datasets = raw_datasets.map(
                tokenize_function,
                batched=True,
                num_proc=data_args.preprocessing_num_workers,
                remove_columns=["text"],
                load_from_cache_file=not data_args.overwrite_cache,
                desc="Running tokenizer on dataset line_by_line",
            )

            train_dataset = tokenized_datasets["train"]
            eval_dataset = tokenized_datasets["test"]

    # Data collator
    # This one will take care of randomly masking the tokens.
    pad_to_multiple_of_8 = training_args.fp16 and not data_args.pad_to_max_length
    data_collator = DataCollatorForLanguageModeling(
        tokenizer=tokenizer,
        mlm_probability=data_args.mlm_probability,
        pad_to_multiple_of=8 if pad_to_multiple_of_8 else None,
    )

    # Initialize our Trainer
    trainer = Trainer(
        model=model,
        args=training_args,
        train_dataset=train_dataset if training_args.do_train else None,
        eval_dataset=eval_dataset if training_args.do_eval else None,
        tokenizer=tokenizer,
        data_collator=data_collator,
    )

    # Training
    if training_args.do_train:
        checkpoint = None
        if training_args.resume_from_checkpoint is not None:
            checkpoint = training_args.resume_from_checkpoint
        elif last_checkpoint is not None:
            checkpoint = last_checkpoint
        train_result = trainer.train(resume_from_checkpoint=checkpoint)
        trainer.save_model()  # Saves the tokenizer too for easy upload
        metrics = train_result.metrics


        try:
            max_train_samples = (
                data_args.max_train_samples if data_args.max_train_samples is not None else len(train_dataset)
            )
            metrics["train_samples"] = min(max_train_samples, len(train_dataset))
        except TypeError:
            # TPU =>
            metrics["train_samples"] = None

        trainer.log_metrics("train", metrics)
        trainer.save_metrics("train", metrics)
        trainer.save_state()

    # Evaluation
    if training_args.do_eval:
        logger.info("*** Evaluate ***")

        metrics = trainer.evaluate()

        try:
            max_eval_samples = data_args.max_eval_samples if data_args.max_eval_samples is not None else len(eval_dataset)
            metrics["eval_samples"] = min(max_eval_samples, len(eval_dataset))
        except:
            metrics["train_samples"] = None

        try:
            perplexity = math.exp(metrics["eval_loss"])
        except OverflowError:
            perplexity = float("inf")
        metrics["perplexity"] = perplexity

        trainer.log_metrics("eval", metrics)
        trainer.save_metrics("eval", metrics)

    if training_args.push_to_hub:
        kwargs = {"finetuned_from": model_args.model_name_or_path, "tasks": "fill-mask"}
        if data_args.dataset_name is not None:
            kwargs["dataset_tags"] = data_args.dataset_name
            if data_args.dataset_config_name is not None:
                kwargs["dataset_args"] = data_args.dataset_config_name
                kwargs["dataset"] = f"{data_args.dataset_name} {data_args.dataset_config_name}"
            else:
                kwargs["dataset"] = data_args.dataset_name

        trainer.push_to_hub(**kwargs)


def _mp_fn(index):
    # For xla_spawn (TPUs)
    main()


if __name__ == "__main__":
    main()
	"""
	This is a workaround for `examples/run_mlm.py` for pretraining models
	with big text files line-by-line.

	For the time being, `datasets` is facing some issues dealing with really
	big text files, so we use a custom dataset until this is fixed.

	August 3th 2021

	Author: Juan Manuel Pérez

	"""


	#!/usr/bin/env python
	# coding=utf-8
	# Copyright 2020 The HuggingFace Team All rights reserved.
	#
	# Licensed under the Apache License, Version 2.0 (the "License");
	# you may not use this file except in compliance with the License.
	# You may obtain a copy of the License at
	#
	# http://www.apache.org/licenses/LICENSE-2.0
	#
	# Unless required by applicable law or agreed to in writing, software
	# distributed under the License is distributed on an "AS IS" BASIS,
	# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
	# See the License for the specific language governing permissions and
	# limitations under the License.
	"""
	Fine-tuning the library models for masked language modeling (BERT, ALBERT, RoBERTa...) on a text file or a dataset.

	Here is the full list of checkpoints on the hub that can be fine-tuned by this script:
	https://huggingface.co/models?filter=masked-lm
	"""
	# You can also adapt this script on your own masked language modeling task. Pointers for this are left as comments.

	import logging
	import math
	from multiprocessing.sharedctypes import Value
	import os
	import sys
	from dataclasses import dataclass, field
	from typing import Optional
	from glob import glob
	import datasets
	from datasets import load_dataset
	import random
	import transformers
	from transformers import (
	CONFIG_MAPPING,
	MODEL_FOR_MASKED_LM_MAPPING,
	AutoConfig,
	AutoModelForMaskedLM,
	AutoTokenizer,
	DataCollatorForLanguageModeling,
	HfArgumentParser,
	Trainer,
	TrainingArguments,
	set_seed,
	)
	from transformers.trainer_utils import get_last_checkpoint
	from transformers.utils import check_min_version
	from transformers.utils.versions import require_version
	from torch.utils.data import IterableDataset, Dataset

	# Will error if the minimal version of Transformers is not installed. Remove at your own risks.
	check_min_version("4.9.0")

	require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/language-modeling/requirements.txt")

	logger = logging.getLogger(__name__)
	MODEL_CONFIG_CLASSES = list(MODEL_FOR_MASKED_LM_MAPPING.keys())
	MODEL_TYPES = tuple(conf.model_type for conf in MODEL_CONFIG_CLASSES)


	class BatchProcessedDataset(IterableDataset):
	"""
	A dataset which streams and process tweets from files
	"""
	def __init__(self, files, tokenizer, batch_size=4096, limit=-1, padding='max_length'):
	"""
	Constructor

	Arguments:
	----------


	files: list[str]
	A list of files from which lines are streamed in order

	tokenizer:
	A transformer's tokenizer

	batch_size:
	The batch size of lines that is passed to the tokenizer

	limit: int (optional)
	An optional limit to the number of streamed lines. By default no limit is set

	padding:

	"""
	self.files = files
	self.batch_size = batch_size
	self.tokenizer = tokenizer
	self.padding = padding
	self.limit = limit

	def __nextbatch(self, f):
	nb = [x.strip("\n") for _, x in zip(range(self.batch_size), f)]
	return nb

	def __iter__(self):
	num_iter = 0
	for file_path in self.files:
	logger.info(f"Opening file {file_path}")
	with open(file_path) as f:
	next_batch = self.__nextbatch(f)
	while next_batch:
	tokenized_batch = self.tokenizer(next_batch, padding=self.padding, truncation=True, return_special_tokens_mask=True)
	for encoding in tokenized_batch.encodings:
	if num_iter > 0 and num_iter >= self.limit:
	return
	yield {
	"input_ids": encoding.ids,
	"token_type_ids": encoding.type_ids,
	"attention_mask": encoding.attention_mask,
	"special_tokens_mask": encoding.special_tokens_mask
	}
	num_iter += 1
	next_batch = self.__nextbatch(f)


	@dataclass
	class ModelArguments:
	"""
	Arguments pertaining to which model/config/tokenizer we are going to fine-tune, or train from scratch.
	"""

	model_name_or_path: Optional[str] = field(
	default=None,
	metadata={
	"help": "The model checkpoint for weights initialization."
	"Don't set if you want to train a model from scratch."
	},
	)
	model_type: Optional[str] = field(
	default=None,
	metadata={"help": "If training from scratch, pass a model type from the list: " + ", ".join(MODEL_TYPES)},
	)
	config_overrides: Optional[str] = field(
	default=None,
	metadata={
	"help": "Override some existing default config settings when a model is trained from scratch. Example: "
	"n_embd=10,resid_pdrop=0.2,scale_attn_weights=false,summary_type=cls_index"
	},
	)
	config_name: Optional[str] = field(
	default=None, metadata={"help": "Pretrained config name or path if not the same as model_name"}
	)
	tokenizer_name: Optional[str] = field(
	default=None, metadata={"help": "Pretrained tokenizer name or path if not the same as model_name"}
	)



	cache_dir: Optional[str] = field(
	default=None,
	metadata={"help": "Where do you want to store the pretrained models downloaded from huggingface.co"},
	)
	use_fast_tokenizer: bool = field(
	default=True,
	metadata={"help": "Whether to use one of the fast tokenizer (backed by the tokenizers library) or not."},
	)
	model_revision: str = field(
	default="main",
	metadata={"help": "The specific model version to use (can be a branch name, tag name or commit id)."},
	)
	use_auth_token: bool = field(
	default=False,
	metadata={
	"help": "Will use the token generated when running `transformers-cli login` (necessary to use this script "
	"with private models)."
	},
	)



	def __post_init__(self):
	if self.config_overrides is not None and (self.config_name is not None or self.model_name_or_path is not None):
	raise ValueError(
	"--config_overrides can't be used in combination with --config_name or --model_name_or_path"
	)


	@dataclass
	class DataTrainingArguments:
	"""
	Arguments pertaining to what data we are going to input our model for training and eval.
	"""

	train_dir: Optional[str] = field(
	default=None,
	metadata={"help": "Path to train files."}
	)
	eval_dir: Optional[str] = field(
	default=None,
	metadata={"help": "Path to test files."}
	)

	max_seq_length: Optional[int] = field(
	default=None,
	metadata={
	"help": "The maximum total input sequence length after tokenization. Sequences longer "
	"than this will be truncated."
	},
	)
	preprocessing_num_workers: Optional[int] = field(
	default=None,
	metadata={"help": "The number of processes to use for the preprocessing."},
	)
	mlm_probability: float = field(
	default=0.15, metadata={"help": "Ratio of tokens to mask for masked language modeling loss"}
	)

	pad_to_max_length: bool = field(
	default=False,
	metadata={
	"help": "Whether to pad all samples to `max_seq_length`. "
	"If False, will pad the samples dynamically when batching to the maximum length in the batch."
	},
	)
	max_train_samples: Optional[int] = field(
	default=None,
	metadata={
	"help": "For debugging purposes or quicker training, truncate the number of training examples to this "
	"value if set."
	},
	)
	max_eval_samples: Optional[int] = field(
	default=None,
	metadata={
	"help": "For debugging purposes or quicker training, truncate the number of evaluation examples to this "
	"value if set."
	},
	)

	tokenization_batch_size: Optional[int] = field(
	default=4096, metadata={
	"help": "Tokenization batch size (for on-the-fly tokenization)"
	}
	)

	tokenize_on_the_fly: bool = field(
	default=False, metadata={
	"help": "Tokenize on-the-fly"
	}
	)

	def __post_init__(self):
	if (self.train_dir is None or self.eval_dir is None:
	raise ValueError("Must provide train_dir & eval_dir")


	def main():
	# See all possible arguments in src/transformers/training_args.py
	# or by passing the --help flag to this script.
	# We now keep distinct sets of args, for a cleaner separation of concerns.

	parser = HfArgumentParser((ModelArguments, DataTrainingArguments, TrainingArguments))
	if (2 <= len(sys.argv) <= 4):
	if sys.argv[-1].endswith(".json"):
	json_file = sys.argv[-1]
	else:
	json_file = sys.argv[1]
	# If we pass only one argument to the script and it's the path to a json file,
	# let's parse it to get our arguments.
	model_args, data_args, training_args = parser.parse_json_file(json_file=os.path.abspath(json_file))
	else:
	model_args, data_args, training_args = parser.parse_args_into_dataclasses()

	print("Model args: ",model_args)
	# Setup logging
	logging.basicConfig(
	format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",
	datefmt="%m/%d/%Y %H:%M:%S",
	handlers=[logging.StreamHandler(sys.stdout)],
	)

	log_level = training_args.get_process_log_level()
	logger.setLevel(log_level)
	datasets.utils.logging.set_verbosity(log_level)
	transformers.utils.logging.set_verbosity(log_level)
	transformers.utils.logging.enable_default_handler()
	transformers.utils.logging.enable_explicit_format()

	# Log on each process the small summary:
	logger.warning(
	f"Process rank: {training_args.local_rank}, device: {training_args.device}, n_gpu: {training_args.n_gpu}"
	+ f"distributed training: {bool(training_args.local_rank != -1)}, 16-bits training: {training_args.fp16}"
	)
	# Set the verbosity to info of the Transformers logger (on main process only):
	logger.info(f"Training/evaluation parameters {training_args}")

	# Detecting last checkpoint.
	last_checkpoint = None
	if os.path.isdir(training_args.output_dir) and training_args.do_train and not training_args.overwrite_output_dir:
	last_checkpoint = get_last_checkpoint(training_args.output_dir)
	if last_checkpoint is None and len(os.listdir(training_args.output_dir)) > 0:
	raise ValueError(
	f"Output directory ({training_args.output_dir}) already exists and is not empty. "
	"Use --overwrite_output_dir to overcome."
	)
	elif last_checkpoint is not None and training_args.resume_from_checkpoint is None:
	logger.info(
	f"Checkpoint detected, resuming training at {last_checkpoint}. To avoid this behavior, change "
	"the `--output_dir` or add `--overwrite_output_dir` to train from scratch."
	)

	# Set seed before initializing model.
	set_seed(training_args.seed)

	config_kwargs = {
	"cache_dir": model_args.cache_dir,
	"revision": model_args.model_revision,
	"use_auth_token": True if model_args.use_auth_token else None,
	}
	if model_args.config_name:
	config = AutoConfig.from_pretrained(model_args.config_name, **config_kwargs)
	elif model_args.model_name_or_path:
	config = AutoConfig.from_pretrained(model_args.model_name_or_path, **config_kwargs)
	else:
	config = CONFIG_MAPPING[model_args.model_type]()
	logger.warning("You are instantiating a new config instance from scratch.")
	if model_args.config_overrides is not None:
	logger.info(f"Overriding config: {model_args.config_overrides}")
	config.update_from_string(model_args.config_overrides)

	tokenizer_kwargs = {
	"cache_dir": model_args.cache_dir,
	"use_fast": model_args.use_fast_tokenizer,
	"revision": model_args.model_revision,
	"use_auth_token": True if model_args.use_auth_token else None,
	}
	if model_args.tokenizer_name:
	tokenizer = AutoTokenizer.from_pretrained(model_args.tokenizer_name, **tokenizer_kwargs)
	elif model_args.model_name_or_path:
	tokenizer = AutoTokenizer.from_pretrained(model_args.model_name_or_path, **tokenizer_kwargs)
	else:
	raise ValueError(
	"You are instantiating a new tokenizer from scratch. This is not supported by this script."
	"You can do it from another script, save it, and load it from here, using --tokenizer_name."
	)

	if model_args.model_name_or_path:
	model = AutoModelForMaskedLM.from_pretrained(
	model_args.model_name_or_path,
	from_tf=bool(".ckpt" in model_args.model_name_or_path),
	config=config,
	cache_dir=model_args.cache_dir,
	revision=model_args.model_revision,
	use_auth_token=True if model_args.use_auth_token else None,
	)
	else:
	logger.info("Training new model from scratch")
	model = AutoModelForMaskedLM.from_config(config)

	model.resize_token_embeddings(len(tokenizer))
	tokenizer.model_max_length =128
	# Get the datasets: you can either provide your own CSV/JSON/TXT training and evaluation files (see below)
	# or just provide the name of one of the public datasets available on the hub at https://huggingface.co/datasets/
	# (the dataset will be downloaded automatically from the datasets Hub
	#
	# For CSV/JSON files, this script will use the column called 'text' or the first column. You can easily tweak this
	# behavior (see below)
	#
	# In distributed training, the load_dataset function guarantee that only one local process can concurrently
	# download the dataset.
	padding = "max_length" if data_args.pad_to_max_length else False

	train_files = glob(os.path.join(data_args.train_dir, "*.txt"))
	eval_files = glob(os.path.join(data_args.eval_dir, "*.txt"))

	if data_args.tokenize_on_the_fly:
	"""
	Use our custom class
	"""
	random.shuffle(train_files)

	if data_args.max_eval_samples is None or not type(data_args.max_eval_samples) is int:
	raise ValueError("Must provide max_eval_samples")
	logger.info(f"Tokenization batch size {data_args.tokenization_batch_size}")
	train_dataset = BatchProcessedDataset(
	train_files, tokenizer, batch_size=data_args.tokenization_batch_size,
	padding=padding,
	)
	eval_dataset = BatchProcessedDataset(
	eval_files, tokenizer, batch_size=data_args.tokenization_batch_size,
	padding=padding, limit=data_args.max_eval_samples
	)

	else:
	"""
	Load using datasets
	"""

	raw_datasets = load_dataset("text",
	data_files = {"train": train_files, "test": eval_files}
	)


	if data_args.max_seq_length is None:
	max_seq_length = tokenizer.model_max_length
	if max_seq_length > 1024:
	logger.warning(
	f"The tokenizer picked seems to have a very large `model_max_length` ({tokenizer.model_max_length}). "
	"Picking 1024 instead. You can change that default value by passing --max_seq_length xxx."
	)
	max_seq_length = 1024
	else:
	if data_args.max_seq_length > tokenizer.model_max_length:
	logger.warning(
	f"The max_seq_length passed ({data_args.max_seq_length}) is larger than the maximum length for the"
	f"model ({tokenizer.model_max_length}). Using max_seq_length={tokenizer.model_max_length}."
	)
	max_seq_length = min(data_args.max_seq_length, tokenizer.model_max_length)

	if data_args.tokenize_on_the_fly:
	training_args.remove_unused_columns = False
	else:
	def tokenize_function(examples):
	# Remove empty lines
	examples["text"] = [
	line for line in examples["text"] if len(line) > 0 and not line.isspace()
	]
	return tokenizer(
	examples["text"],
	padding=padding,
	truncation=True,
	max_length=max_seq_length,
	# We use this option because DataCollatorForLanguageModeling (see below) is more efficient when it
	# receives the `special_tokens_mask`.
	return_special_tokens_mask=True,
	)

	with training_args.main_process_first(desc="dataset map tokenization"):
	tokenized_datasets = raw_datasets.map(
	tokenize_function,
	batched=True,
	num_proc=data_args.preprocessing_num_workers,
	remove_columns=["text"],
	load_from_cache_file=not data_args.overwrite_cache,
	desc="Running tokenizer on dataset line_by_line",
	)

	train_dataset = tokenized_datasets["train"]
	eval_dataset = tokenized_datasets["test"]

	# Data collator
	# This one will take care of randomly masking the tokens.
	pad_to_multiple_of_8 = training_args.fp16 and not data_args.pad_to_max_length
	data_collator = DataCollatorForLanguageModeling(
	tokenizer=tokenizer,
	mlm_probability=data_args.mlm_probability,
	pad_to_multiple_of=8 if pad_to_multiple_of_8 else None,
	)

	# Initialize our Trainer
	trainer = Trainer(
	model=model,
	args=training_args,
	train_dataset=train_dataset if training_args.do_train else None,
	eval_dataset=eval_dataset if training_args.do_eval else None,
	tokenizer=tokenizer,
	data_collator=data_collator,
	)

	# Training
	if training_args.do_train:
	checkpoint = None
	if training_args.resume_from_checkpoint is not None:
	checkpoint = training_args.resume_from_checkpoint
	elif last_checkpoint is not None:
	checkpoint = last_checkpoint
	train_result = trainer.train(resume_from_checkpoint=checkpoint)
	trainer.save_model() # Saves the tokenizer too for easy upload
	metrics = train_result.metrics


	try:
	max_train_samples = (
	data_args.max_train_samples if data_args.max_train_samples is not None else len(train_dataset)
	)
	metrics["train_samples"] = min(max_train_samples, len(train_dataset))
	except TypeError:
	# TPU =>
	metrics["train_samples"] = None

	trainer.log_metrics("train", metrics)
	trainer.save_metrics("train", metrics)
	trainer.save_state()

	# Evaluation
	if training_args.do_eval:
	logger.info("* Evaluate *")

	metrics = trainer.evaluate()

	try:
	max_eval_samples = data_args.max_eval_samples if data_args.max_eval_samples is not None else len(eval_dataset)
	metrics["eval_samples"] = min(max_eval_samples, len(eval_dataset))
	except:
	metrics["train_samples"] = None

	try:
	perplexity = math.exp(metrics["eval_loss"])
	except OverflowError:
	perplexity = float("inf")
	metrics["perplexity"] = perplexity

	trainer.log_metrics("eval", metrics)
	trainer.save_metrics("eval", metrics)

	if training_args.push_to_hub:
	kwargs = {"finetuned_from": model_args.model_name_or_path, "tasks": "fill-mask"}
	if data_args.dataset_name is not None:
	kwargs["dataset_tags"] = data_args.dataset_name
	if data_args.dataset_config_name is not None:
	kwargs["dataset_args"] = data_args.dataset_config_name
	kwargs["dataset"] = f"{data_args.dataset_name} {data_args.dataset_config_name}"
	else:
	kwargs["dataset"] = data_args.dataset_name

	trainer.push_to_hub(**kwargs)


	def _mp_fn(index):
	# For xla_spawn (TPUs)
	main()


	if __name__ == "__main__":
	main()