train bert for token classification

convlab2/base_models/bert/create_data.py

+import os
+import json
+from tqdm import tqdm
+from convlab2.util import load_dataset, load_nlu_data, load_dst_data, load_policy_data, load_nlg_data, load_e2e_data, load_rg_data
+from nltk.tokenize import TreebankWordTokenizer, PunktSentenceTokenizer
+from collections import Counter
+
+def create_bio_data(dataset, data_dir):
+    data_by_split = load_nlu_data(dataset, speaker='all')
+    os.makedirs(data_dir, exist_ok=True)
+
+    sent_tokenizer = PunktSentenceTokenizer()
+    word_tokenizer = TreebankWordTokenizer()
+
+    data_splits = data_by_split.keys()
+    cnt = Counter()
+    for data_split in data_splits:
+        data = []
+        for sample in tqdm(data_by_split[data_split], desc=f'{data_split} sample', leave=False):
+            utterance = sample['utterance']
+            dialogue_acts = [da for da in sample['dialogue_acts']['non-categorical'] if 'start' in da]
+            cnt[len(dialogue_acts)] += 1
+
+            sentences = sent_tokenizer.tokenize(utterance)
+            sent_spans = sent_tokenizer.span_tokenize(utterance)
+            tokens = [token for sent in sentences for token in word_tokenizer.tokenize(sent)]
+            token_spans = [(sent_span[0]+token_span[0], sent_span[0]+token_span[1]) for sent, sent_span in zip(sentences, sent_spans) for token_span in word_tokenizer.span_tokenize(sent)]
+            labels = ['O'] * len(tokens)
+            for da in dialogue_acts:
+                char_start = da['start']
+                char_end = da['end']
+                word_start, word_end = -1, -1
+                for i, token_span in enumerate(token_spans):
+                    if char_start == token_span[0]:
+                        word_start = i
+                    if char_end == token_span[1]:
+                        word_end = i + 1
+                if word_start == -1 and word_end == -1:
+                    # char span does not match word, skip
+                    continue
+                labels[word_start] = 'B'
+                for i in range(word_start+1, word_end):
+                    labels[i] = "I"
+            data.append(json.dumps({'tokens': tokens, 'labels': labels})+'\n')
+        file_name = os.path.join(data_dir, f"{data_split}.json")
+        with open(file_name, "w") as f:
+            f.writelines(data)
+    print('num of spans in utterances', cnt)
+
+if __name__ == '__main__':
+    from argparse import ArgumentParser
+    parser = ArgumentParser(description="create data for seq2seq training")
+    parser.add_argument('--tasks', metavar='task_name', nargs='*', choices=['bio'], help='names of tasks')
+    parser.add_argument('--datasets', metavar='dataset_name', nargs='*', help='names of unified datasets')
+    parser.add_argument('--save_dir', metavar='save_directory', type=str, default='data', help='directory to save the data, default: data/$task_name/$dataset_name')
+    args = parser.parse_args()
+    print(args)
+    for dataset_name in tqdm(args.datasets, desc='datasets'):
+        dataset = load_dataset(dataset_name)
+        for task_name in tqdm(args.tasks, desc='tasks', leave=False):
+            data_dir = os.path.join(args.save_dir, task_name, dataset_name)
+            eval(f"create_{task_name}_data")(dataset, data_dir)

convlab2/base_models/bert/run_bio.sh

+n_gpus=8
+task_name="bio"
+dataset_name="sgd"
+data_dir="data/${task_name}/${dataset_name}"
+output_dir="output/${task_name}/${dataset_name}"
+cache_dir="cache"
+logging_dir="${output_dir}/runs"
+train_file="${data_dir}/train.json"
+validation_file="${data_dir}/validation.json"
+test_file="${data_dir}/test.json"
+source_column="tokens"
+target_column="labels"
+model_name_or_path="bert-base-uncased"
+per_device_train_batch_size=128
+per_device_eval_batch_size=512
+gradient_accumulation_steps=1
+lr=2e-5
+num_train_epochs=1
+metric_for_best_model="f1"
+
+python create_data.py --tasks ${task_name} --datasets ${dataset_name} --save_dir "data"
+
+python -m torch.distributed.launch \
+    --nproc_per_node ${n_gpus} run_token_classification.py \
+    --task_name ${task_name} \
+    --train_file ${train_file} \
+    --validation_file ${validation_file} \
+    --test_file ${test_file} \
+    --source_column ${source_column} \
+    --target_column ${target_column} \
+    --model_name_or_path ${model_name_or_path} \
+    --do_train \
+    --do_eval \
+    --do_predict \
+    --save_strategy epoch \
+    --evaluation_strategy epoch \
+    --load_best_model_at_end \
+    --metric_for_best_model ${metric_for_best_model} \
+    --cache_dir ${cache_dir} \
+    --output_dir ${output_dir} \
+    --logging_dir ${logging_dir} \
+    --overwrite_output_dir \
+    --preprocessing_num_workers 4 \
+    --per_device_train_batch_size ${per_device_train_batch_size} \
+    --per_device_eval_batch_size ${per_device_eval_batch_size} \
+    --gradient_accumulation_steps ${gradient_accumulation_steps} \
+    --learning_rate ${lr} \
+    --num_train_epochs ${num_train_epochs} \
+    --debug underflow_overflow

convlab2/base_models/bert/run_token_classification.py

+#!/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 token classification.
+Modified from https://github.com/huggingface/transformers/blob/master/examples/pytorch/token-classification/run_ner.py
+"""
+# You can also adapt this script on your own token classification task and datasets. Pointers for this are left as comments.
+
+import logging
+import os
+import sys
+import json
+from dataclasses import dataclass, field
+from typing import Optional
+
+import datasets
+import numpy as np
+from datasets import ClassLabel, load_dataset, load_metric
+
+import transformers
+from transformers import (
+    AutoConfig,
+    AutoModelForTokenClassification,
+    AutoTokenizer,
+    DataCollatorForTokenClassification,
+    HfArgumentParser,
+    PreTrainedTokenizerFast,
+    Trainer,
+    TrainingArguments,
+    set_seed,
+)
+from transformers.trainer_utils import EvalPrediction, get_last_checkpoint
+from transformers.utils import check_min_version
+from transformers.utils.versions import require_version
+
+
+# Will error if the minimal version of Transformers is not installed. Remove at your own risks.
+check_min_version("4.12.5")
+
+require_version("datasets>=1.16.1")
+
+logger = logging.getLogger(__name__)
+os.environ["WANDB_DISABLED"] = "true"
+
+
+@dataclass
+class ModelArguments:
+    """
+    Arguments pertaining to which model/config/tokenizer we are going to fine-tune from.
+    """
+
+    model_name_or_path: str = field(
+        metadata={"help": "Path to pretrained model or model identifier from huggingface.co/models"}
+    )
+    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 to store the pretrained models downloaded from huggingface.co"},
+    )
+    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)."
+        },
+    )
+
+
+@dataclass
+class DataTrainingArguments:
+    """
+    Arguments pertaining to what data we are going to input our model for training and eval.
+    """
+
+
+    task_name: Optional[str] = field(
+        default=None, metadata={"help": "The name of the task (ner, pos...)."}
+    )
+    dataset_name: Optional[str] = field(
+        default=None, metadata={"help": "The name of the dataset to use (via the datasets library)."}
+    )
+    dataset_config_name: Optional[str] = field(
+        default=None, metadata={"help": "The configuration name of the dataset to use (via the datasets library)."}
+    )
+    source_column: Optional[str] = field(
+        default=None,
+        metadata={"help": "The name of the column in the datasets containing the source texts."},
+    )
+    target_column: Optional[str] = field(
+        default=None,
+        metadata={"help": "The name of the column in the datasets containing the target labels."},
+    )
+    train_file: Optional[str] = field(
+        default=None, metadata={"help": "The input training data file (a jsonlines or csv file)."}
+    )
+    validation_file: Optional[str] = field(
+        default=None,
+        metadata={
+            "help": "An optional input evaluation data file to evaluate the metrics on (a jsonlines or csv file)."
+        },
+    )
+    test_file: Optional[str] = field(
+        default=None,
+        metadata={
+            "help": "An optional input test data file to evaluate the metrics on (a jsonlines or csv file)."
+        },
+    )
+    overwrite_cache: bool = field(
+        default=False, metadata={"help": "Overwrite the cached training and evaluation sets"}
+    )
+    preprocessing_num_workers: Optional[int] = field(
+        default=None,
+        metadata={"help": "The number of processes to use for the preprocessing."},
+    )
+    max_seq_length: int = field(
+        default=None,
+        metadata={
+            "help": "The maximum total input sequence length after tokenization. If set, sequences longer "
+            "than this will be truncated, sequences shorter will be padded."
+        },
+    )
+    pad_to_max_length: bool = field(
+        default=False,
+        metadata={
+            "help": "Whether to pad all samples to model maximum sentence length. "
+            "If False, will pad the samples dynamically when batching to the maximum length in the batch. More "
+            "efficient on GPU but very bad for TPU."
+        },
+    )
+    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."
+        },
+    )
+    max_predict_samples: Optional[int] = field(
+        default=None,
+        metadata={
+            "help": "For debugging purposes or quicker training, truncate the number of prediction examples to this "
+            "value if set."
+        },
+    )
+    label_all_tokens: bool = field(
+        default=False,
+        metadata={
+            "help": "Whether to put the label for one word on all tokens of generated by that word or just on the "
+            "one (in which case the other tokens will have a padding index)."
+        },
+    )
+    return_entity_level_metrics: bool = field(
+        default=False,
+        metadata={"help": "Whether to return all the entity levels during evaluation or just the overall ones."},
+    )
+
+    def __post_init__(self):
+        if (
+            self.dataset_name is None
+            and self.train_file is None
+            and self.validation_file is None
+            and self.test_file is None
+        ):
+            raise ValueError("Need either a dataset name or a training/validation/testing file.")
+        else:
+            if self.train_file is not None:
+                extension = self.train_file.split(".")[-1]
+                assert extension in ["csv", "json"], "`train_file` should be a csv or a json file."
+            if self.validation_file is not None:
+                extension = self.validation_file.split(".")[-1]
+                assert extension in ["csv", "json"], "`validation_file` should be a csv or a json file."
+            if self.test_file is not None:
+                extension = self.test_file.split(".")[-1]
+                assert extension in ["csv", "json"], "`test_file` should be a csv or a json file."
+
+
+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 len(sys.argv) == 2 and sys.argv[1].endswith(".json"):
+        # 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(sys.argv[1]))
+    else:
+        model_args, data_args, training_args = parser.parse_args_into_dataclasses()
+
+    # 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}"
+    )
+    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)
+
+    # Get the datasets: you can either provide your own CSV/JSON 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 first column for the source texts and the second column for the
+    # target labels (unless you specify column names for this with the `source_column` and `target_column` arguments).
+    #
+    # In distributed training, the load_dataset function guarantee that only one local process can concurrently
+    # download the dataset.
+    if data_args.dataset_name is not None:
+        # Downloading and loading a dataset from the hub.
+        raw_datasets = load_dataset(
+            data_args.dataset_name, data_args.dataset_config_name, cache_dir=model_args.cache_dir
+        )
+    else:
+        data_files = {}
+        if data_args.train_file is not None:
+            data_files["train"] = data_args.train_file
+            extension = data_args.train_file.split(".")[-1]
+        if data_args.validation_file is not None:
+            data_files["validation"] = data_args.validation_file
+            extension = data_args.validation_file.split(".")[-1]
+        if data_args.test_file is not None:
+            data_files["test"] = data_args.test_file
+            extension = data_args.test_file.split(".")[-1]
+        raw_datasets = load_dataset(extension, data_files=data_files, cache_dir=model_args.cache_dir)
+    # See more about loading any type of standard or custom dataset (from files, python dict, pandas DataFrame, etc) at
+    # https://huggingface.co/docs/datasets/loading_datasets.html.
+
+    # must provide train dataset for label space
+    column_names = raw_datasets["train"].column_names
+    features = raw_datasets["train"].features
+
+    if data_args.source_column is None:
+        source_column = column_names[0]
+    else:
+        source_column = data_args.source_column
+        if source_column not in column_names:
+            raise ValueError(
+                f"--source_column' value '{data_args.source_column}' needs to be one of: {', '.join(column_names)}"
+            )
+
+    if data_args.target_column is None:
+        target_column = column_names[1]
+    else:
+        target_column = data_args.target_column
+        if target_column not in column_names:
+            raise ValueError(
+                f"--target_column' value '{data_args.target_column}' needs to be one of: {', '.join(column_names)}"
+            )
+
+    # In the event the labels are not a `Sequence[ClassLabel]`, we will need to go through the dataset to get the
+    # unique labels.
+    def get_label_list(labels):
+        unique_labels = set()
+        for label in labels:
+            unique_labels = unique_labels | set(label)
+        label_list = list(unique_labels)
+        label_list.sort()
+        return label_list
+
+    if isinstance(features[target_column].feature, ClassLabel):
+        label_list = features[target_column].feature.names
+        # No need to convert the labels since they are already ints.
+        label_to_id = {i: i for i in range(len(label_list))}
+    else:
+        label_list = get_label_list(raw_datasets["train"][target_column])
+        label_to_id = {l: i for i, l in enumerate(label_list)}
+    num_labels = len(label_list)
+
+    # Map that sends B-Xxx label to its I-Xxx counterpart for label_all_tokens
+    b_to_i_label = []
+    for idx, label in enumerate(label_list):
+        if label.startswith("B-") and label.replace("B-", "I-") in label_list:
+            b_to_i_label.append(label_list.index(label.replace("B-", "I-")))
+        else:
+            b_to_i_label.append(idx)
+
+    # Load pretrained model and tokenizer
+    #
+    # Distributed training:
+    # The .from_pretrained methods guarantee that only one local process can concurrently
+    # download model & vocab.
+    config = AutoConfig.from_pretrained(
+        model_args.config_name if model_args.config_name else model_args.model_name_or_path,
+        num_labels=num_labels,
+        label2id=label_to_id,
+        id2label={i: l for l, i in label_to_id.items()},
+        finetuning_task=data_args.task_name,
+        cache_dir=model_args.cache_dir,
+        revision=model_args.model_revision,
+        use_auth_token=True if model_args.use_auth_token else None,
+    )
+
+    tokenizer_name_or_path = model_args.tokenizer_name if model_args.tokenizer_name else model_args.model_name_or_path
+    if config.model_type in {"gpt2", "roberta"}:
+        tokenizer = AutoTokenizer.from_pretrained(
+            tokenizer_name_or_path,
+            cache_dir=model_args.cache_dir,
+            use_fast=True,
+            revision=model_args.model_revision,
+            use_auth_token=True if model_args.use_auth_token else None,
+            add_prefix_space=True,
+        )
+    else:
+        tokenizer = AutoTokenizer.from_pretrained(
+            tokenizer_name_or_path,
+            cache_dir=model_args.cache_dir,
+            use_fast=True,
+            revision=model_args.model_revision,
+            use_auth_token=True if model_args.use_auth_token else None,
+        )
+
+    model = AutoModelForTokenClassification.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,
+    )
+
+    # Tokenizer check: this script requires a fast tokenizer.
+    if not isinstance(tokenizer, PreTrainedTokenizerFast):
+        raise ValueError(
+            "This example script only works for models that have a fast tokenizer. Checkout the big table of models "
+            "at https://huggingface.co/transformers/index.html#supported-frameworks to find the model types that meet this "
+            "requirement"
+        )
+
+    # Preprocessing the dataset
+    # Padding strategy
+    padding = "max_length" if data_args.pad_to_max_length else False
+
+    # Tokenize all texts and align the labels with them.
+    def tokenize_and_align_labels(examples):
+        tokenized_inputs = tokenizer(
+            examples[source_column],
+            padding=padding,
+            truncation=True,
+            max_length=data_args.max_seq_length,
+            # We use this argument because the texts in our dataset are lists of words (with a label for each word).
+            is_split_into_words=True,
+        )
+        labels = []
+        for i, label in enumerate(examples[target_column]):
+            word_ids = tokenized_inputs.word_ids(batch_index=i)
+            previous_word_idx = None
+            label_ids = []
+            for word_idx in word_ids:
+                # Special tokens have a word id that is None. We set the label to -100 so they are automatically
+                # ignored in the loss function.
+                if word_idx is None:
+                    label_ids.append(-100)
+                # We set the label for the first token of each word.
+                elif word_idx != previous_word_idx:
+                    label_ids.append(label_to_id[label[word_idx]])
+                # For the other tokens in a word, we set the label to either the current label or -100, depending on
+                # the label_all_tokens flag.
+                else:
+                    if data_args.label_all_tokens:
+                        label_ids.append(b_to_i_label[label_to_id[label[word_idx]]])
+                    else:
+                        label_ids.append(-100)
+                previous_word_idx = word_idx
+
+            labels.append(label_ids)
+        tokenized_inputs["labels"] = labels
+        return tokenized_inputs
+
+    if training_args.do_train:
+        if "train" not in raw_datasets:
+            raise ValueError("--do_train requires a train dataset")
+        train_dataset = raw_datasets["train"]
+        if data_args.max_train_samples is not None:
+            train_dataset = train_dataset.select(range(data_args.max_train_samples))
+        with training_args.main_process_first(desc="train dataset map pre-processing"):
+            train_dataset = train_dataset.map(
+                tokenize_and_align_labels,
+                batched=True,
+                num_proc=data_args.preprocessing_num_workers,
+                load_from_cache_file=not data_args.overwrite_cache,
+                desc="Running tokenizer on train dataset",
+            )
+
+    if training_args.do_eval:
+        if "validation" not in raw_datasets:
+            raise ValueError("--do_eval requires a validation dataset")
+        eval_dataset = raw_datasets["validation"]
+        if data_args.max_eval_samples is not None:
+            eval_dataset = eval_dataset.select(range(data_args.max_eval_samples))
+        with training_args.main_process_first(desc="validation dataset map pre-processing"):
+            eval_dataset = eval_dataset.map(
+                tokenize_and_align_labels,
+                batched=True,
+                num_proc=data_args.preprocessing_num_workers,
+                load_from_cache_file=not data_args.overwrite_cache,
+                desc="Running tokenizer on validation dataset",
+            )
+
+    if training_args.do_predict:
+        if "test" not in raw_datasets:
+            raise ValueError("--do_predict requires a test dataset")
+        predict_dataset = raw_datasets["test"]
+        if data_args.max_predict_samples is not None:
+            predict_dataset = predict_dataset.select(range(data_args.max_predict_samples))
+        with training_args.main_process_first(desc="prediction dataset map pre-processing"):
+            predict_dataset = predict_dataset.map(
+                tokenize_and_align_labels,
+                batched=True,
+                num_proc=data_args.preprocessing_num_workers,
+                load_from_cache_file=not data_args.overwrite_cache,
+                desc="Running tokenizer on prediction dataset",
+            )
+
+    # Data collator
+    data_collator = DataCollatorForTokenClassification(tokenizer, pad_to_multiple_of=8 if training_args.fp16 else None)
+
+    # Metrics
+    metric = load_metric("seqeval")
+
+    def compute_metrics(p: EvalPrediction):
+        predictions, labels = p
+        predictions = np.argmax(predictions, axis=2)
+
+        # Remove ignored index (special tokens)
+        true_predictions = [
+            [label_list[p] for (p, l) in zip(prediction, label) if l != -100]
+            for prediction, label in zip(predictions, labels)
+        ]
+        true_labels = [
+            [label_list[l] for (p, l) in zip(prediction, label) if l != -100]
+            for prediction, label in zip(predictions, labels)
+        ]
+
+        results = metric.compute(predictions=true_predictions, references=true_labels)
+        if data_args.return_entity_level_metrics:
+            # Unpack nested dictionaries
+            final_results = {}
+            for key, value in results.items():
+                if isinstance(value, dict):
+                    for n, v in value.items():
+                        final_results[f"{key}_{n}"] = v
+                else:
+                    final_results[key] = value
+            return final_results
+        else:
+            return {
+                "precision": results["overall_precision"],
+                "recall": results["overall_recall"],
+                "f1": results["overall_f1"],
+                "accuracy": results["overall_accuracy"],
+            }
+
+    # 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,
+        compute_metrics=compute_metrics,
+    )
+
+    # 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
+        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))
+
+        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()
+
+        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))
+
+        trainer.log_metrics("eval", metrics)
+        trainer.save_metrics("eval", metrics)
+
+    # Predict
+    if training_args.do_predict:
+        logger.info("*** Predict ***")
+
+        predictions, labels, metrics = trainer.predict(predict_dataset, metric_key_prefix="predict")
+        predictions = np.argmax(predictions, axis=2)
+
+        # Remove ignored index (special tokens)
+        true_predictions = [
+            [label_list[p] for (p, l) in zip(prediction, label) if l != -100]
+            for prediction, label in zip(predictions, labels)
+        ]
+
+        trainer.log_metrics("predict", metrics)
+        trainer.save_metrics("predict", metrics)
+
+        # Save predictions
+        output_predictions_file = os.path.join(training_args.output_dir, "predictions.json")
+        if trainer.is_world_process_zero():
+            with open(output_predictions_file, "w") as writer:
+                for sample, prediction in zip(raw_datasets["test"], true_predictions):
+                    sample["predictions"] = prediction
+                    writer.write(json.dumps(sample)+'\n')
+
+    kwargs = {"finetuned_from": model_args.model_name_or_path, "tasks": data_args.task_name}
+    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
+
+    if training_args.push_to_hub:
+        trainer.push_to_hub(**kwargs)
+    else:
+        trainer.create_model_card(**kwargs)
+
+
+def _mp_fn(index):
+    # For xla_spawn (TPUs)
+    main()
+
+
+if __name__ == "__main__":
+    main()

setup.py

@@ -42,6 +42,7 @@ setup(
         'torch>=1.6',
         'transformers>=4.0',
         'datasets>=1.8',
+        'seqeval',
         'spacy',
         'allennlp',
         'simplejson',