Merge pull request #1548 from huggingface/cli

[2.2] - Command-line interface - Pipeline class

transformers/modeling_utils.py

@@ -281,7 +281,9 @@ class PreTrainedModel(nn.Module):
             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:`~transformers.PretrainedConfig`:
+            config: (`optional`) one of:
+                    - an instance of a class derived from :class:`~transformers.PretrainedConfig`, or
+                    - a string valid as input to :func:`~transformers.PretrainedConfig.from_pretrained()`
                 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
@@ -336,10 +338,11 @@ class PreTrainedModel(nn.Module):
         proxies = kwargs.pop('proxies', None)
         output_loading_info = kwargs.pop('output_loading_info', False)
 
-        # Load config
-        if config is None:
+        # Load config if we don't provide a configuration
+        if not isinstance(config, PretrainedConfig):
+            config_path = config if config is not None else pretrained_model_name_or_path
             config, model_kwargs = cls.config_class.from_pretrained(
-                pretrained_model_name_or_path, *model_args,
+                config_path, *model_args,
                 cache_dir=cache_dir, return_unused_kwargs=True,
                 force_download=force_download,
                 resume_download=resume_download,
@@ -408,7 +411,11 @@ class PreTrainedModel(nn.Module):
         model = cls(config, *model_args, **model_kwargs)
 
         if state_dict is None and not from_tf:
+            try:
                 state_dict = torch.load(resolved_archive_file, map_location='cpu')
+            except:
+                raise OSError("Unable to load weights from pytorch checkpoint file. "
+                            "If you tried to load a PyTorch model from a TF 2.0 checkpoint, please set from_tf=True. ")
 
         missing_keys = []
         unexpected_keys = []

transformers/pipelines.py

+# coding=utf-8
+# Copyright 2018 The HuggingFace Inc. team.
+#
+# 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.
+from __future__ import absolute_import, division, print_function, unicode_literals
+
+import csv
+import json
+import os
+import pickle
+import logging
+import six
+from abc import ABC, abstractmethod
+from contextlib import contextmanager
+from itertools import groupby
+from os.path import abspath, exists
+from typing import Union, Optional, Tuple, List, Dict
+
+import numpy as np
+
+from transformers import (AutoConfig, AutoTokenizer, PreTrainedTokenizer,
+                          PretrainedConfig, ModelCard, SquadExample,
+                          squad_convert_examples_to_features, is_tf_available,
+                          is_torch_available, BasicTokenizer,
+                          ALL_PRETRAINED_CONFIG_ARCHIVE_MAP)
+
+if is_tf_available():
+    import tensorflow as tf
+    from transformers import TFAutoModel, TFAutoModelForSequenceClassification, \
+        TFAutoModelForQuestionAnswering, TFAutoModelForTokenClassification
+
+if is_torch_available():
+    import torch
+    from transformers import AutoModel, AutoModelForSequenceClassification, \
+        AutoModelForQuestionAnswering, AutoModelForTokenClassification
+
+
+logger = logging.getLogger(__name__)
+
+def get_framework(model=None):
+    """ Select framework (TensorFlow/PyTorch) to use.
+        If both frameworks are installed and no specific model is provided, defaults to using TensorFlow.
+    """
+    if is_tf_available() and is_torch_available() and model is not None and not isinstance(model, str):
+        # Both framework are available but the use supplied a model class instance.
+        # Try to guess which framework to use from the model classname
+        framework = 'tf' if model.__class__.__name__.startswith('TF') else 'pt'
+    elif not is_tf_available() and not is_torch_available():
+        raise ImportError("At least one of TensorFlow 2.0 or PyTorch should be installed. "
+                          "To install TensorFlow 2.0, read the instructions at https://www.tensorflow.org/install/ "
+                          "To install PyTorch, read the instructions at https://pytorch.org/.")
+    else:
+        framework = 'tf' if is_tf_available() else 'pt'
+    return framework
+
+class ArgumentHandler(ABC):
+    """
+    Base interface for handling varargs for each Pipeline
+    """
+    @abstractmethod
+    def __call__(self, *args, **kwargs):
+        raise NotImplementedError()
+
+
+class DefaultArgumentHandler(ArgumentHandler):
+    """
+    Default varargs argument parser handling parameters for each Pipeline
+    """
+    def __call__(self, *args, **kwargs):
+        if 'X' in kwargs:
+            return kwargs['X']
+        elif 'data' in kwargs:
+            return kwargs['data']
+        elif len(args) == 1:
+            if isinstance(args[0], list):
+                return args[0]
+            else:
+                return [args[0]]
+        elif len(args) > 1:
+            return list(args)
+        raise ValueError('Unable to infer the format of the provided data (X=, data=, ...)')
+
+
+class PipelineDataFormat:
+    """
+    Base class for all the pipeline supported data format both for reading and writing.
+    Supported data formats currently includes:
+     - JSON
+     - CSV
+
+    PipelineDataFormat also includes some utilities to work with multi-columns like mapping from datasets columns
+    to pipelines keyword arguments through the `dataset_kwarg_1=dataset_column_1` format.
+    """
+    SUPPORTED_FORMATS = ['json', 'csv', 'pipe']
+
+    def __init__(self, output: Optional[str], input: Optional[str], column: Optional[str]):
+        self.output = output
+        self.path = input
+        self.column = column.split(',') if column else ['']
+        self.is_multi_columns = len(self.column) > 1
+
+        if self.is_multi_columns:
+            self.column = [tuple(c.split('=')) if '=' in c else (c, c) for c in self.column]
+
+        if output is not None:
+            if exists(abspath(self.output)):
+                raise OSError('{} already exists on disk'.format(self.output))
+
+        if input is not None:
+            if not exists(abspath(self.path)):
+                raise OSError('{} doesnt exist on disk'.format(self.path))
+
+    @abstractmethod
+    def __iter__(self):
+        raise NotImplementedError()
+
+    @abstractmethod
+    def save(self, data: dict):
+        """
+        Save the provided data object with the representation for the current `DataFormat`.
+        :param data: data to store
+        :return:
+        """
+        raise NotImplementedError()
+
+    def save_binary(self, data: Union[dict, List[dict]]) -> str:
+        """
+        Save the provided data object as a pickle-formatted binary data on the disk.
+        :param data: data to store
+        :return: (str) Path where the data has been saved
+        """
+        path, _ = os.path.splitext(self.output)
+        binary_path = os.path.extsep.join((path, 'pickle'))
+
+        with open(binary_path, 'wb+') as f_output:
+            pickle.dump(data, f_output)
+
+        return binary_path
+
+    @staticmethod
+    def from_str(name: str, output: Optional[str], path: Optional[str], column: Optional[str]):
+        if name == 'json':
+            return JsonPipelineDataFormat(output, path, column)
+        elif name == 'csv':
+            return CsvPipelineDataFormat(output, path, column)
+        elif name == 'pipe':
+            return PipedPipelineDataFormat(output, path, column)
+        else:
+            raise KeyError('Unknown reader {} (Available reader are json/csv/pipe)'.format(name))
+
+
+class CsvPipelineDataFormat(PipelineDataFormat):
+    def __init__(self, output: Optional[str], input: Optional[str], column: Optional[str]):
+        super().__init__(output, input, column)
+
+    def __iter__(self):
+        with open(self.path, 'r') as f:
+            reader = csv.DictReader(f)
+            for row in reader:
+                if self.is_multi_columns:
+                    yield {k: row[c] for k, c in self.column}
+                else:
+                    yield row[self.column[0]]
+
+    def save(self, data: List[dict]):
+        with open(self.output, 'w') as f:
+            if len(data) > 0:
+                writer = csv.DictWriter(f, list(data[0].keys()))
+                writer.writeheader()
+                writer.writerows(data)
+
+
+class JsonPipelineDataFormat(PipelineDataFormat):
+    def __init__(self, output: Optional[str], input: Optional[str], column: Optional[str]):
+        super().__init__(output, input, column)
+
+        with open(input, 'r') as f:
+            self._entries = json.load(f)
+
+    def __iter__(self):
+        for entry in self._entries:
+            if self.is_multi_columns:
+                yield {k: entry[c] for k, c in self.column}
+            else:
+                yield entry[self.column[0]]
+
+    def save(self, data: dict):
+        with open(self.output, 'w') as f:
+            json.dump(data, f)
+
+
+class PipedPipelineDataFormat(PipelineDataFormat):
+    """
+    Read data from piped input to the python process.
+    For multi columns data, columns should separated by \t
+
+    If columns are provided, then the output will be a dictionary with {column_x: value_x}
+    """
+    def __iter__(self):
+        import sys
+        for line in sys.stdin:
+
+            # Split for multi-columns
+            if '\t' in line:
+
+                line = line.split('\t')
+                if self.column:
+                    # Dictionary to map arguments
+                    yield {kwargs: l for (kwargs, _), l in zip(self.column, line)}
+                else:
+                    yield tuple(line)
+
+            # No dictionary to map arguments
+            else:
+                yield line
+
+    def save(self, data: dict):
+        print(data)
+
+    def save_binary(self, data: Union[dict, List[dict]]) -> str:
+        if self.output is None:
+            raise KeyError(
+                'When using piped input on pipeline outputting large object requires an output file path. '
+                'Please provide such output path through --output argument.'
+            )
+
+        return super().save_binary(data)
+
+
+class _ScikitCompat(ABC):
+    """
+    Interface layer for the Scikit and Keras compatibility.
+    """
+
+    @abstractmethod
+    def transform(self, X):
+        raise NotImplementedError()
+
+    @abstractmethod
+    def predict(self, X):
+        raise NotImplementedError()
+
+
+class Pipeline(_ScikitCompat):
+    """
+    Base class implementing pipelined operations.
+    Pipeline workflow is defined as a sequence of the following operations:
+        Input -> Tokenization -> Model Inference -> Post-Processing (Task dependent) -> Output
+
+    Pipeline supports running on CPU or GPU through the device argument. Users can specify
+    device argument as an integer, -1 meaning "CPU", >= 0 referring the CUDA device ordinal.
+
+    Some pipeline, like for instance FeatureExtractionPipeline ('feature-extraction') outputs large
+    tensor object as nested-lists. In order to avoid dumping such large structure as textual data we
+    provide the binary_output constructor argument. If set to True, the output will be stored in the
+    pickle format.
+
+    Arguments:
+        **model**: ``(str, PretrainedModel, TFPretrainedModel)``:
+            Reference to the model to use through this pipeline.
+
+        **tokenizer**: ``(str, PreTrainedTokenizer)``:
+            Reference to the tokenizer to use through this pipeline.
+
+        **args_parser**: ``ArgumentHandler``:
+            Reference to the object in charge of parsing supplied pipeline parameters.
+
+        **device**: ``int``:
+            Device ordinal for CPU/GPU supports. Setting this to -1 will leverage CPU, >=0 will run the model
+            on the associated CUDA device id.
+
+        **binary_output** ``bool`` (default: False):
+            Flag indicating if the output the pipeline should happen in a binary format (i.e. pickle) or as raw text.
+
+    Return:
+        Pipeline returns list or dictionary depending on:
+         - Does the user provided multiple sample
+         - The pipeline expose multiple fields in the output object
+
+    Examples:
+        nlp = pipeline('ner')
+        nlp = pipeline('ner', model='...', config='...', tokenizer='...')
+        nlp = NerPipeline(model='...', config='...', tokenizer='...')
+        nlp = QuestionAnsweringPipeline(model=AutoModel.from_pretrained('...'), tokenizer='...')
+    """
+    def __init__(self, model, tokenizer: PreTrainedTokenizer = None,
+                 modelcard: ModelCard = None, framework: Optional[str] = None,
+                 args_parser: ArgumentHandler = None, device: int = -1,
+                 binary_output: bool = False):
+
+        if framework is None:
+            framework = get_framework()
+
+        self.model = model
+        self.tokenizer = tokenizer
+        self.modelcard = modelcard
+        self.framework = framework
+        self.device = device
+        self.binary_output = binary_output
+        self._args_parser = args_parser or DefaultArgumentHandler()
+
+        # Special handling
+        if self.device >= 0 and self.framework == 'pt':
+            self.model = self.model.to('cuda:{}'.format(self.device))
+
+    def save_pretrained(self, save_directory):
+        """
+        Save the pipeline's model and tokenizer to the specified save_directory
+        """
+        if not os.path.isdir(save_directory):
+            logger.error("Provided path ({}) should be a directory".format(save_directory))
+            return
+
+        self.model.save_pretrained(save_directory)
+        self.tokenizer.save_pretrained(save_directory)
+        self.modelcard.save_pretrained(save_directory)
+
+    def transform(self, X):
+        """
+        Scikit / Keras interface to transformers' pipelines. This method will forward to __call__().
+        """
+        return self(X=X)
+
+    def predict(self, X):
+        """
+        Scikit / Keras interface to transformers' pipelines. This method will forward to __call__().
+        Se
+        """
+        return self(X=X)
+
+    @contextmanager
+    def device_placement(self):
+        """
+        Context Manager allowing tensor allocation on the user-specified device in framework agnostic way.
+        example:
+            # Explicitly ask for tensor allocation on CUDA device :0
+            nlp = pipeline(..., device=0)
+            with nlp.device_placement():
+                # Every framework specific tensor allocation will be done on the request device
+                output = nlp(...)
+        Returns:
+            Context manager
+        """
+        if self.framework == 'tf':
+            with tf.device('/CPU:0' if self.device == -1 else '/device:GPU:{}'.format(self.device)):
+                yield
+        else:
+            if self.device >= 0:
+                torch.cuda.set_device(self.device)
+
+            yield
+
+    def inputs_for_model(self, features: Union[dict, List[dict]]) -> Dict:
+        """
+        Generates the input dictionary with model-specific parameters.
+
+        Returns:
+            dict holding all the required parameters for model's forward
+        """
+        args = ['input_ids', 'attention_mask']
+        model_type = type(self.model).__name__.lower()
+
+        if 'distilbert' not in model_type and 'xlm' not in model_type:
+            args += ['token_type_ids']
+
+        # PR #1548 (CLI) There is an issue with attention_mask
+        # if 'xlnet' in model_type or 'xlm' in model_type:
+        #     args += ['cls_index', 'p_mask']
+
+        if isinstance(features, dict):
+            return {k: features[k] for k in args}
+        else:
+            return {k: [feature[k] for feature in features] for k in args}
+
+    def __call__(self, *texts, **kwargs):
+        # Parse arguments
+        inputs = self._args_parser(*texts, **kwargs)
+
+        # Encode for forward
+        with self.device_placement():
+            inputs = self.tokenizer.batch_encode_plus(
+                inputs, add_special_tokens=True,
+                return_tensors=self.framework,
+                max_length=self.tokenizer.max_len
+            )
+
+            # Filter out features not available on specific models
+            inputs = self.inputs_for_model(inputs)
+            return self._forward(inputs)
+
+    def _forward(self, inputs):
+        """
+        Internal framework specific forward dispatching.
+        Args:
+            inputs: dict holding all the keyworded arguments for required by the model forward method.
+        Returns:
+            Numpy array
+        """
+        if self.framework == 'tf':
+            # TODO trace model
+            predictions = self.model(inputs, training=False)[0]
+        else:
+            with torch.no_grad():
+                predictions = self.model(**inputs)[0].cpu()
+
+        return predictions.numpy()
+
+
+class FeatureExtractionPipeline(Pipeline):
+    """
+    Feature extraction pipeline using Model head.
+    """
+
+    def __init__(self, model,
+                 tokenizer: PreTrainedTokenizer = None,
+                 modelcard: ModelCard = None,
+                 framework: Optional[str] = None,
+                 args_parser: ArgumentHandler = None,
+                 device: int = -1):
+        super().__init__(model=model,
+                         tokenizer=tokenizer,
+                         modelcard=modelcard,
+                         framework=framework,
+                         args_parser=args_parser,
+                         device=device,
+                         binary_output=True)
+
+    def __call__(self, *args, **kwargs):
+        return super().__call__(*args, **kwargs).tolist()
+
+
+class TextClassificationPipeline(Pipeline):
+    """
+    Text classification pipeline using ModelForTextClassification head.
+    """
+
+    def __call__(self, *args, **kwargs):
+        outputs = super().__call__(*args, **kwargs)
+        scores = np.exp(outputs) / np.exp(outputs).sum(-1)
+        return [{'label': self.model.config.id2label[item.argmax()], 'score': item.max()} for item in scores]
+
+
+class NerPipeline(Pipeline):
+    """
+    Named Entity Recognition pipeline using ModelForTokenClassification head.
+    """
+
+    def __init__(self, model, tokenizer: PreTrainedTokenizer = None,
+                 modelcard: ModelCard = None, framework: Optional[str] = None,
+                 args_parser: ArgumentHandler = None, device: int = -1,
+                 binary_output: bool = False):
+        super().__init__(model=model,
+                         tokenizer=tokenizer,
+                         modelcard=modelcard,
+                         framework=framework,
+                         args_parser=args_parser,
+                         device=device,
+                         binary_output=binary_output)
+
+        self._basic_tokenizer = BasicTokenizer(do_lower_case=False)
+
+    def __call__(self, *texts, **kwargs):
+        inputs, answers = self._args_parser(*texts, **kwargs), []
+        for sentence in inputs:
+
+            # Ugly token to word idx mapping (for now)
+            token_to_word, words = [], self._basic_tokenizer.tokenize(sentence)
+            for i, w in enumerate(words):
+                tokens = self.tokenizer.tokenize(w)
+                token_to_word += [i] * len(tokens)
+
+            # Manage correct placement of the tensors
+            with self.device_placement():
+
+                tokens = self.tokenizer.encode_plus(
+                    sentence, return_attention_mask=False,
+                    return_tensors=self.framework,
+                    max_length=self.tokenizer.max_len
+                )
+
+                # Forward
+                if self.framework == 'tf':
+                    entities = self.model(tokens)[0][0].numpy()
+                else:
+                    with torch.no_grad():
+                        entities = self.model(**tokens)[0][0].cpu().numpy()
+
+            # Normalize scores
+            answer, token_start = [], 1
+            for idx, word in groupby(token_to_word):
+
+                # Sum log prob over token, then normalize across labels
+                score = np.exp(entities[token_start]) / np.exp(entities[token_start]).sum(-1, keepdims=True)
+                label_idx = score.argmax()
+
+                if label_idx > 0:
+                    answer += [{
+                        'word': words[idx],
+                        'score': score[label_idx].item(),
+                        'entity': self.model.config.id2label[label_idx]
+                    }]
+
+                # Update token start
+                token_start += len(list(word))
+
+            # Append
+            answers += [answer]
+        return answers
+
+
+class QuestionAnsweringArgumentHandler(ArgumentHandler):
+    """
+    QuestionAnsweringPipeline requires the user to provide multiple arguments (i.e. question & context) to be mapped
+    to internal SquadExample / SquadFeature structures.
+
+    QuestionAnsweringArgumentHandler manages all the possible to create SquadExample from the command-line supplied
+    arguments.
+    """
+    def __call__(self, *args, **kwargs):
+        # Position args, handling is sensibly the same as X and data, so forwarding to avoid duplicating
+        if args is not None and len(args) > 0:
+            if len(args) == 1:
+                kwargs['X'] = args[0]
+            else:
+                kwargs['X'] = list(args)
+
+        # Generic compatibility with sklearn and Keras
+        # Batched data
+        if 'X' in kwargs or 'data' in kwargs:
+            inputs = kwargs['X'] if 'X' in kwargs else kwargs['data']
+
+            if isinstance(inputs, dict):
+                inputs = [inputs]
+            else:
+                # Copy to avoid overriding arguments
+                inputs = [i for i in inputs]
+
+            for i, item in enumerate(inputs):
+                if isinstance(item, dict):
+                    if any(k not in item for k in ['question', 'context']):
+                        raise KeyError('You need to provide a dictionary with keys {question:..., context:...}')
+
+                    inputs[i] = QuestionAnsweringPipeline.create_sample(**item)
+
+                elif not isinstance(item, SquadExample):
+                    raise ValueError(
+                        '{} argument needs to be of type (list[SquadExample | dict], SquadExample, dict)'
+                            .format('X' if 'X' in kwargs else 'data')
+                    )
+
+            # Tabular input
+        elif 'question' in kwargs and 'context' in kwargs:
+            if isinstance(kwargs['question'], str):
+                kwargs['question'] = [kwargs['question']]
+
+            if isinstance(kwargs['context'], str):
+                kwargs['context'] = [kwargs['context']]
+
+            inputs = [QuestionAnsweringPipeline.create_sample(q, c) for q, c in zip(kwargs['question'], kwargs['context'])]
+        else:
+            raise ValueError('Unknown arguments {}'.format(kwargs))
+
+        if not isinstance(inputs, list):
+            inputs = [inputs]
+
+        return inputs
+
+
+class QuestionAnsweringPipeline(Pipeline):
+    """
+    Question Answering pipeline using ModelForQuestionAnswering head.
+    """
+
+    def __init__(self, model,
+                 tokenizer: Optional[PreTrainedTokenizer],
+                 modelcard: Optional[ModelCard],
+                 framework: Optional[str] = None,
+                 device: int = -1, **kwargs):
+        super().__init__(model=model,
+                         tokenizer=tokenizer,
+                         modelcard=modelcard,
+                         framework=framework,
+                         args_parser=QuestionAnsweringArgumentHandler(),
+                         device=device, **kwargs)
+
+    @staticmethod
+    def create_sample(question: Union[str, List[str]], context: Union[str, List[str]]) -> Union[SquadExample, List[SquadExample]]:
+        """
+        QuestionAnsweringPipeline leverages the SquadExample/SquadFeatures internally.
+        This helper method encapsulate all the logic for converting question(s) and context(s) to SquadExample(s).
+        We currently support extractive question answering.
+        Arguments:
+             question: (str, List[str]) The question to be ask for the associated context
+             context: (str, List[str]) The context in which we will look for the answer.
+
+        Returns:
+            SquadExample initialized with the corresponding question and context.
+        """
+        if isinstance(question, list):
+            return [SquadExample(None, q, c, None, None, None) for q, c in zip(question, context)]
+        else:
+            return SquadExample(None, question, context, None, None, None)
+
+    def __call__(self, *texts, **kwargs):
+        """
+        Args:
+            We support multiple use-cases, the following are exclusive:
+            X: sequence of SquadExample
+            data: sequence of SquadExample
+            question: (str, List[str]), batch of question(s) to map along with context
+            context: (str, List[str]), batch of context(s) associated with the provided question keyword argument
+        Returns:
+            dict: {'answer': str, 'score": float, 'start": int, "end": int}
+            answer: the textual answer in the intial context
+            score: the score the current answer scored for the model
+            start: the character index in the original string corresponding to the beginning of the answer' span
+            end: the character index in the original string corresponding to the ending of the answer' span
+        """
+        # Set defaults values
+        kwargs.setdefault('topk', 1)
+        kwargs.setdefault('doc_stride', 128)
+        kwargs.setdefault('max_answer_len', 15)
+        kwargs.setdefault('max_seq_len', 384)
+        kwargs.setdefault('max_question_len', 64)
+
+        if kwargs['topk'] < 1:
+            raise ValueError('topk parameter should be >= 1 (got {})'.format(kwargs['topk']))
+
+        if kwargs['max_answer_len'] < 1:
+            raise ValueError('max_answer_len parameter should be >= 1 (got {})'.format(kwargs['max_answer_len']))
+
+        # Convert inputs to features
+        examples = self._args_parser(*texts, **kwargs)
+        features = squad_convert_examples_to_features(examples, self.tokenizer, kwargs['max_seq_len'], kwargs['doc_stride'], kwargs['max_question_len'], False)
+        fw_args = self.inputs_for_model([f.__dict__ for f in features])
+
+        # Manage tensor allocation on correct device
+        with self.device_placement():
+            if self.framework == 'tf':
+                fw_args = {k: tf.constant(v) for (k, v) in fw_args.items()}
+                start, end = self.model(fw_args)
+                start, end = start.numpy(), end.numpy()
+            else:
+                with torch.no_grad():
+                    # Retrieve the score for the context tokens only (removing question tokens)
+                    fw_args = {k: torch.tensor(v) for (k, v) in fw_args.items()}
+                    start, end = self.model(**fw_args)
+                    start, end = start.cpu().numpy(), end.cpu().numpy()
+
+        answers = []
+        for (example, feature, start_, end_) in zip(examples, features, start, end):
+            # Normalize logits and spans to retrieve the answer
+            start_ = np.exp(start_) / np.sum(np.exp(start_))
+            end_ = np.exp(end_) / np.sum(np.exp(end_))
+
+            # Mask padding and question
+            start_, end_ = start_ * np.abs(np.array(feature.p_mask) - 1), end_ * np.abs(np.array(feature.p_mask) - 1)
+
+            # TODO : What happens if not possible
+            # Mask CLS
+            start_[0] = end_[0] = 0
+
+            starts, ends, scores = self.decode(start_, end_, kwargs['topk'], kwargs['max_answer_len'])
+            char_to_word = np.array(example.char_to_word_offset)
+
+            # Convert the answer (tokens) back to the original text
+            answers += [
+                {
+                    'score': score.item(),
+                    'start': np.where(char_to_word == feature.token_to_orig_map[s])[0][0].item(),
+                    'end': np.where(char_to_word == feature.token_to_orig_map[e])[0][-1].item(),
+                    'answer': ' '.join(example.doc_tokens[feature.token_to_orig_map[s]:feature.token_to_orig_map[e] + 1])
+                }
+                for s, e, score in zip(starts, ends, scores)
+            ]
+
+        if len(answers) == 1:
+            return answers[0]
+        return answers
+
+    def decode(self, start: np.ndarray, end: np.ndarray, topk: int, max_answer_len: int) -> Tuple:
+        """
+        Take the output of any QuestionAnswering head and will generate probalities for each span to be
+        the actual answer.
+        In addition, it filters out some unwanted/impossible cases like answer len being greater than
+        max_answer_len or answer end position being before the starting position.
+        The method supports output the k-best answer through the topk argument.
+
+        Args:
+            start: numpy array, holding individual start probabilities for each token
+            end: numpy array, holding individual end probabilities for each token
+            topk: int, indicates how many possible answer span(s) to extract from the model's output
+            max_answer_len: int, maximum size of the answer to extract from the model's output
+        """
+        # Ensure we have batch axis
+        if start.ndim == 1:
+            start = start[None]
+
+        if end.ndim == 1:
+            end = end[None]
+
+        # Compute the score of each tuple(start, end) to be the real answer
+        outer = np.matmul(np.expand_dims(start, -1), np.expand_dims(end, 1))
+
+        # Remove candidate with end < start and end - start > max_answer_len
+        candidates = np.tril(np.triu(outer), max_answer_len - 1)
+
+        #  Inspired by Chen & al. (https://github.com/facebookresearch/DrQA)
+        scores_flat = candidates.flatten()
+        if topk == 1:
+            idx_sort = [np.argmax(scores_flat)]
+        elif len(scores_flat) < topk:
+            idx_sort = np.argsort(-scores_flat)
+        else:
+            idx = np.argpartition(-scores_flat, topk)[0:topk]
+            idx_sort = idx[np.argsort(-scores_flat[idx])]
+
+        start, end = np.unravel_index(idx_sort, candidates.shape)[1:]
+        return start, end, candidates[0, start, end]
+
+    def span_to_answer(self, text: str, start: int, end: int):
+        """
+        When decoding from token probalities, this method maps token indexes to actual word in
+        the initial context.
+
+        Args:
+            text: str, the actual context to extract the answer from
+            start: int, starting answer token index
+            end: int, ending answer token index
+
+        Returns:
+            dict: {'answer': str, 'start': int, 'end': int}
+        """
+        words = []
+        token_idx = char_start_idx = char_end_idx = chars_idx = 0
+
+        for i, word in enumerate(text.split(" ")):
+            token = self.tokenizer.tokenize(word)
+
+            # Append words if they are in the span
+            if start <= token_idx <= end:
+                if token_idx == start:
+                    char_start_idx = chars_idx
+
+                if token_idx == end:
+                    char_end_idx = chars_idx + len(word)
+
+                words += [word]
+
+            # Stop if we went over the end of the answer
+            if token_idx > end:
+                break
+
+            # Append the subtokenization length to the running index
+            token_idx += len(token)
+            chars_idx += len(word) + 1
+
+        # Join text with spaces
+        return {'answer': ' '.join(words), 'start': max(0, char_start_idx), 'end': min(len(text), char_end_idx)}
+
+
+# Register all the supported task here
+SUPPORTED_TASKS = {
+    'feature-extraction': {
+        'impl': FeatureExtractionPipeline,
+        'tf': TFAutoModel if is_tf_available() else None,
+        'pt': AutoModel if is_torch_available() else None,
+        'default': {
+            'model': {
+                'pt': 'distilbert-base-uncased',
+                'tf': 'distilbert-base-uncased',
+            },
+            'config': None,
+            'tokenizer': 'distilbert-base-uncased'
+        }
+    },
+    'sentiment-analysis': {
+        'impl': TextClassificationPipeline,
+        'tf': TFAutoModelForSequenceClassification if is_tf_available() else None,
+        'pt': AutoModelForSequenceClassification if is_torch_available() else None,
+        'default': {
+            'model': {
+                'pt': 'https://s3.amazonaws.com/models.huggingface.co/bert/distilbert-base-uncased-finetuned-sst-2-english-pytorch_model.bin',
+                'tf': 'https://s3.amazonaws.com/models.huggingface.co/bert/distilbert-base-uncased-finetuned-sst-2-english-tf_model.h5',
+            },
+            'config': 'https://s3.amazonaws.com/models.huggingface.co/bert/distilbert-base-uncased-finetuned-sst-2-english-config.json',
+            'tokenizer': 'distilbert-base-uncased'
+        }
+    },
+    'ner': {
+        'impl': NerPipeline,
+        'tf': TFAutoModelForTokenClassification if is_tf_available() else None,
+        'pt': AutoModelForTokenClassification if is_torch_available() else None,
+        'default': {
+            'model': {
+                'pt':'https://s3.amazonaws.com/models.huggingface.co/bert/bert-large-cased-finetuned-conll03-english-pytorch_model.bin',
+                'tf': 'https://s3.amazonaws.com/models.huggingface.co/bert/bert-large-cased-finetuned-conll03-english-tf_model.h5',
+            },
+            'config': 'https://s3.amazonaws.com/models.huggingface.co/bert/bert-large-cased-finetuned-conll03-english-config.json',
+            'tokenizer': 'bert-large-cased'
+        }
+    },
+    'question-answering': {
+        'impl': QuestionAnsweringPipeline,
+        'tf': TFAutoModelForQuestionAnswering if is_tf_available() else None,
+        'pt': AutoModelForQuestionAnswering if is_torch_available() else None,
+        'default': {
+            'model': {
+                'pt': 'distilbert-base-uncased-distilled-squad',
+                'tf': 'distilbert-base-uncased-distilled-squad',
+            },
+            'config': None,
+            'tokenizer': 'distilbert-base-uncased'
+        }
+    }
+}
+
+
+def pipeline(task: str, model: Optional = None,
+             config: Optional[Union[str, PretrainedConfig]] = None,
+             tokenizer: Optional[Union[str, PreTrainedTokenizer]] = None,
+             modelcard: Optional[Union[str, ModelCard]] = None,
+             **kwargs) -> Pipeline:
+    """
+    Utility factory method to build a pipeline.
+    Pipeline are made of:
+        A Tokenizer instance in charge of mapping raw textual input to token
+        A Model instance
+        Some (optional) post processing for enhancing model's output
+
+    Examples:
+        pipeline('sentiment-analysis')
+        pipeline('question-answering', model='distilbert-base-uncased-distilled-squad', tokenizer='bert-base-cased')
+        pipeline('ner', model=AutoModel.from_pretrained(...), tokenizer=AutoTokenizer.from_pretrained(...)
+        pipeline('ner', model='https://...pytorch-model.bin', config='https://...config.json', tokenizer='bert-base-cased')
+    """
+    # Retrieve the task
+    if task not in SUPPORTED_TASKS:
+        raise KeyError("Unknown task {}, available tasks are {}".format(task, list(SUPPORTED_TASKS.keys())))
+
+    framework = get_framework(model)
+
+    targeted_task = SUPPORTED_TASKS[task]
+    task, model_class = targeted_task['impl'], targeted_task[framework]
+
+    # Use default model/config/tokenizer for the task if no model is provided
+    if model is None:
+        models, config, tokenizer = tuple(targeted_task['default'].values())
+        model = models[framework]
+
+    # Try to infer tokenizer from model or config name (if provided as str)
+    if tokenizer is None:
+        if isinstance(model, str) and model in ALL_PRETRAINED_CONFIG_ARCHIVE_MAP:
+            tokenizer = model
+        elif isinstance(config, str) and config in ALL_PRETRAINED_CONFIG_ARCHIVE_MAP:
+            tokenizer = config
+        else:
+            # Impossible to guest what is the right tokenizer here
+            raise Exception("Impossible to guess which tokenizer to use. "
+                            "Please provided a PretrainedTokenizer class or a path/url/shortcut name to a pretrained tokenizer.")
+
+    # Try to infer modelcard from model or config name (if provided as str)
+    if modelcard is None:
+        # Try to fallback on one of the provided string for model or config (will replace the suffix)
+        if isinstance(model, str):
+            modelcard = model
+        elif isinstance(config, str):
+            modelcard = config
+
+    # Instantiate tokenizer if needed
+    if isinstance(tokenizer, six.string_types):
+        tokenizer = AutoTokenizer.from_pretrained(tokenizer)
+
+    # Instantiate config if needed
+    if isinstance(config, str):
+        config = AutoConfig.from_pretrained(config)
+
+    # Instantiate model if needed
+    if isinstance(model, str):
+        # Handle transparent TF/PT model conversion
+        model_kwargs = {}
+        if framework == 'pt' and model.endswith('.h5'):
+            model_kwargs['from_tf'] = True
+            logger.warning('Model might be a TensorFlow model (ending with `.h5`) but TensorFlow is not available. '
+                           'Trying to load the model with PyTorch.')
+        elif framework == 'tf' and model.endswith('.bin'):
+            model_kwargs['from_pt'] = True
+            logger.warning('Model might be a PyTorch model (ending with `.bin`) but PyTorch is not available. '
+                           'Trying to load the model with Tensorflow.')
+        model = model_class.from_pretrained(model, config=config, **model_kwargs)
+
+    return task(model=model, tokenizer=tokenizer, modelcard=modelcard, framework=framework, **kwargs)

transformers/tests/model_card_test.py

@@ -18,7 +18,7 @@ import os
 import json
 import unittest
 
-from transformers.model_card import ModelCard
+from transformers.modelcard import ModelCard
 from .tokenization_tests_commons import TemporaryDirectory
 
 class ModelCardTester(unittest.TestCase):
@@ -49,20 +49,20 @@ class ModelCardTester(unittest.TestCase):
                             }
 
     def test_model_card_common_properties(self):
-        model_card = ModelCard.from_dict(self.inputs_dict)
-        self.assertTrue(hasattr(model_card, 'model_details'))
-        self.assertTrue(hasattr(model_card, 'intended_use'))
-        self.assertTrue(hasattr(model_card, 'factors'))
-        self.assertTrue(hasattr(model_card, 'metrics'))
-        self.assertTrue(hasattr(model_card, 'evaluation_data'))
-        self.assertTrue(hasattr(model_card, 'training_data'))
-        self.assertTrue(hasattr(model_card, 'quantitative_analyses'))
-        self.assertTrue(hasattr(model_card, 'ethical_considerations'))
-        self.assertTrue(hasattr(model_card, 'caveats_and_recommendations'))
+        modelcard = ModelCard.from_dict(self.inputs_dict)
+        self.assertTrue(hasattr(modelcard, 'model_details'))
+        self.assertTrue(hasattr(modelcard, 'intended_use'))
+        self.assertTrue(hasattr(modelcard, 'factors'))
+        self.assertTrue(hasattr(modelcard, 'metrics'))
+        self.assertTrue(hasattr(modelcard, 'evaluation_data'))
+        self.assertTrue(hasattr(modelcard, 'training_data'))
+        self.assertTrue(hasattr(modelcard, 'quantitative_analyses'))
+        self.assertTrue(hasattr(modelcard, 'ethical_considerations'))
+        self.assertTrue(hasattr(modelcard, 'caveats_and_recommendations'))
 
     def test_model_card_to_json_string(self):
-        model_card = ModelCard.from_dict(self.inputs_dict)
-        obj = json.loads(model_card.to_json_string())
+        modelcard = ModelCard.from_dict(self.inputs_dict)
+        obj = json.loads(modelcard.to_json_string())
         for key, value in self.inputs_dict.items():
             self.assertEqual(obj[key], value)
 
@@ -70,7 +70,7 @@ class ModelCardTester(unittest.TestCase):
         model_card_first = ModelCard.from_dict(self.inputs_dict)
 
         with TemporaryDirectory() as tmpdirname:
-            filename = os.path.join(tmpdirname, u"model_card.json")
+            filename = os.path.join(tmpdirname, u"modelcard.json")
             model_card_first.to_json_file(filename)
             model_card_second = ModelCard.from_json_file(filename)
 

transformers/tests/pipelines_test.py

+import unittest
+
+from typing import Iterable
+
+from transformers import pipeline
+from transformers.tests.utils import require_tf, require_torch
+
+QA_FINETUNED_MODELS = {
+    ('bert-base-uncased', 'bert-large-uncased-whole-word-masking-finetuned-squad', None),
+    ('bert-base-cased', 'bert-large-cased-whole-word-masking-finetuned-squad', None),
+    ('bert-base-uncased', 'distilbert-base-uncased-distilled-squad', None)
+}
+
+TF_QA_FINETUNED_MODELS = {
+    ('bert-base-uncased', 'bert-large-uncased-whole-word-masking-finetuned-squad', None),
+    ('bert-base-cased', 'bert-large-cased-whole-word-masking-finetuned-squad', None),
+    ('bert-base-uncased', 'distilbert-base-uncased-distilled-squad', None)
+}
+
+TF_NER_FINETUNED_MODELS = {
+    (
+        'bert-base-cased',
+        'https://s3.amazonaws.com/models.huggingface.co/bert/bert-large-cased-finetuned-conll03-english-tf_model.h5',
+        'https://s3.amazonaws.com/models.huggingface.co/bert/bert-large-cased-finetuned-conll03-english-config.json'
+    )
+}
+
+NER_FINETUNED_MODELS = {
+    (
+        'bert-base-cased',
+        'https://s3.amazonaws.com/models.huggingface.co/bert/bert-large-cased-finetuned-conll03-english-pytorch_model.bin',
+        'https://s3.amazonaws.com/models.huggingface.co/bert/bert-large-cased-finetuned-conll03-english-config.json'
+    )
+}
+
+FEATURE_EXTRACT_FINETUNED_MODELS = {
+   ('bert-base-cased', 'bert-base-cased', None),
+   # ('xlnet-base-cased', 'xlnet-base-cased', None), # Disabled for now as it crash for TF2
+   ('distilbert-base-uncased', 'distilbert-base-uncased', None)
+}
+
+TF_FEATURE_EXTRACT_FINETUNED_MODELS = {
+   ('bert-base-cased', 'bert-base-cased', None),
+   # ('xlnet-base-cased', 'xlnet-base-cased', None), # Disabled for now as it crash for TF2
+   ('distilbert-base-uncased', 'distilbert-base-uncased', None)
+}
+
+TF_TEXT_CLASSIF_FINETUNED_MODELS = {
+    (
+        'bert-base-uncased',
+        'https://s3.amazonaws.com/models.huggingface.co/bert/distilbert-base-uncased-finetuned-sst-2-english-tf_model.h5',
+        'https://s3.amazonaws.com/models.huggingface.co/bert/distilbert-base-uncased-finetuned-sst-2-english-config.json'
+    )
+}
+
+TEXT_CLASSIF_FINETUNED_MODELS = {
+    (
+        'bert-base-uncased',
+        'https://s3.amazonaws.com/models.huggingface.co/bert/distilbert-base-uncased-finetuned-sst-2-english-pytorch_model.bin',
+        'https://s3.amazonaws.com/models.huggingface.co/bert/distilbert-base-uncased-finetuned-sst-2-english-config.json'
+    )
+}
+
+
+class MonoColumnInputTestCase(unittest.TestCase):
+    def _test_mono_column_pipeline(self, nlp, valid_inputs: list, invalid_inputs: list, output_keys: Iterable[str]):
+        self.assertIsNotNone(nlp)
+
+        mono_result = nlp(valid_inputs[0])
+        self.assertIsInstance(mono_result, list)
+        self.assertIsInstance(mono_result[0], (dict, list))
+
+        if isinstance(mono_result[0], list):
+            mono_result = mono_result[0]
+
+        for key in output_keys:
+            self.assertIn(key, mono_result[0])
+
+        multi_result = nlp(valid_inputs)
+        self.assertIsInstance(multi_result, list)
+        self.assertIsInstance(multi_result[0], (dict, list))
+
+        if isinstance(multi_result[0], list):
+            multi_result = multi_result[0]
+
+        for result in multi_result:
+            for key in output_keys:
+                self.assertIn(key, result)
+
+        self.assertRaises(Exception, nlp, invalid_inputs)
+
+    @require_torch
+    def test_ner(self):
+        mandatory_keys = {'entity', 'word', 'score'}
+        valid_inputs = ['HuggingFace is solving NLP one commit at a time.', 'HuggingFace is based in New-York & Paris']
+        invalid_inputs = [None]
+        for tokenizer, model, config in NER_FINETUNED_MODELS:
+            nlp = pipeline(task='ner', model=model, config=config, tokenizer=tokenizer)
+            self._test_mono_column_pipeline(nlp, valid_inputs, invalid_inputs, mandatory_keys)
+
+    @require_tf
+    def test_tf_ner(self):
+        mandatory_keys = {'entity', 'word', 'score'}
+        valid_inputs = ['HuggingFace is solving NLP one commit at a time.', 'HuggingFace is based in New-York & Paris']
+        invalid_inputs = [None]
+        for tokenizer, model, config in TF_NER_FINETUNED_MODELS:
+            nlp = pipeline(task='ner', model=model, config=config, tokenizer=tokenizer)
+            self._test_mono_column_pipeline(nlp, valid_inputs, invalid_inputs, mandatory_keys)
+
+    @require_torch
+    def test_sentiment_analysis(self):
+        mandatory_keys = {'label'}
+        valid_inputs = ['HuggingFace is solving NLP one commit at a time.', 'HuggingFace is based in New-York & Paris']
+        invalid_inputs = [None]
+        for tokenizer, model, config in TEXT_CLASSIF_FINETUNED_MODELS:
+            nlp = pipeline(task='sentiment-analysis', model=model, config=config, tokenizer=tokenizer)
+            self._test_mono_column_pipeline(nlp, valid_inputs, invalid_inputs, mandatory_keys)
+
+    @require_tf
+    def test_tf_sentiment_analysis(self):
+        mandatory_keys = {'label'}
+        valid_inputs = ['HuggingFace is solving NLP one commit at a time.', 'HuggingFace is based in New-York & Paris']
+        invalid_inputs = [None]
+        for tokenizer, model, config in TF_TEXT_CLASSIF_FINETUNED_MODELS:
+            nlp = pipeline(task='sentiment-analysis', model=model, config=config, tokenizer=tokenizer)
+            self._test_mono_column_pipeline(nlp, valid_inputs, invalid_inputs, mandatory_keys)
+
+    @require_torch
+    def test_features_extraction(self):
+        valid_inputs = ['HuggingFace is solving NLP one commit at a time.', 'HuggingFace is based in New-York & Paris']
+        invalid_inputs = [None]
+        for tokenizer, model, config in FEATURE_EXTRACT_FINETUNED_MODELS:
+            nlp = pipeline(task='sentiment-analysis', model=model, config=config, tokenizer=tokenizer)
+            self._test_mono_column_pipeline(nlp, valid_inputs, invalid_inputs, {})
+
+    @require_tf
+    def test_tf_features_extraction(self):
+        valid_inputs = ['HuggingFace is solving NLP one commit at a time.', 'HuggingFace is based in New-York & Paris']
+        invalid_inputs = [None]
+        for tokenizer, model, config in TF_FEATURE_EXTRACT_FINETUNED_MODELS:
+            nlp = pipeline(task='sentiment-analysis', model=model, config=config, tokenizer=tokenizer)
+            self._test_mono_column_pipeline(nlp, valid_inputs, invalid_inputs, {})
+
+
+class MultiColumnInputTestCase(unittest.TestCase):
+    def _test_multicolumn_pipeline(self, nlp, valid_inputs: list, invalid_inputs: list, output_keys: Iterable[str]):
+        self.assertIsNotNone(nlp)
+
+        mono_result = nlp(valid_inputs[0])
+        self.assertIsInstance(mono_result, dict)
+
+        for key in output_keys:
+            self.assertIn(key, mono_result)
+
+        multi_result = nlp(valid_inputs)
+        self.assertIsInstance(multi_result, list)
+        self.assertIsInstance(multi_result[0], dict)
+
+        for result in multi_result:
+            for key in output_keys:
+                self.assertIn(key, result)
+
+        self.assertRaises(Exception, nlp, invalid_inputs[0])
+        self.assertRaises(Exception, nlp, invalid_inputs)
+
+    @require_torch
+    def test_question_answering(self):
+        mandatory_output_keys = {'score', 'answer', 'start', 'end'}
+        valid_samples = [
+            {'question': 'Where was HuggingFace founded ?', 'context': 'HuggingFace was founded in Paris.'},
+            {
+                'question': 'In what field is HuggingFace working ?',
+                'context': 'HuggingFace is a startup based in New-York founded in Paris which is trying to solve NLP.'
+            }
+        ]
+        invalid_samples = [
+            {'question': '', 'context': 'This is a test to try empty question edge case'},
+            {'question': None, 'context': 'This is a test to try empty question edge case'},
+            {'question': 'What is does with empty context ?', 'context': ''},
+            {'question': 'What is does with empty context ?', 'context': None},
+        ]
+
+        for tokenizer, model, config in QA_FINETUNED_MODELS:
+            nlp = pipeline(task='question-answering', model=model, config=config, tokenizer=tokenizer)
+            self._test_multicolumn_pipeline(nlp, valid_samples, invalid_samples, mandatory_output_keys)
+
+    @require_tf
+    def test_tf_question_answering(self):
+        mandatory_output_keys = {'score', 'answer', 'start', 'end'}
+        valid_samples = [
+            {'question': 'Where was HuggingFace founded ?', 'context': 'HuggingFace was founded in Paris.'},
+            {
+                'question': 'In what field is HuggingFace working ?',
+                'context': 'HuggingFace is a startup based in New-York founded in Paris which is trying to solve NLP.'
+            }
+        ]
+        invalid_samples = [
+            {'question': '', 'context': 'This is a test to try empty question edge case'},
+            {'question': None, 'context': 'This is a test to try empty question edge case'},
+            {'question': 'What is does with empty context ?', 'context': ''},
+            {'question': 'What is does with empty context ?', 'context': None},
+        ]
+
+        for tokenizer, model, config in TF_QA_FINETUNED_MODELS:
+            nlp = pipeline(task='question-answering', model=model, config=config, tokenizer=tokenizer)
+            self._test_multicolumn_pipeline(nlp, valid_samples, invalid_samples, mandatory_output_keys)
+
+
+if __name__ == '__main__':
+    unittest.main()

transformers/tokenization_utils.py

@@ -886,6 +886,92 @@ class PreTrainedTokenizer(object):
                                       return_overflowing_tokens=return_overflowing_tokens,
                                       return_special_tokens_mask=return_special_tokens_mask)
 
+    def batch_encode_plus(self,
+                          batch_text_or_text_pairs=None,
+                          add_special_tokens=False,
+                          max_length=None,
+                          stride=0,
+                          truncation_strategy='longest_first',
+                          return_tensors=None,
+                          return_input_lengths=False,
+                          return_attention_masks=False,
+                          **kwargs):
+        """
+        Returns a dictionary containing the encoded sequence or sequence pair and additional information:
+        the mask for sequence classification and the overflowing elements if a ``max_length`` is specified.
+
+        Args:
+            batch_text_or_text_pairs: Batch of sequences or pair of sequences to be encoded.
+                This can be a list of string/string-sequences/int-sequences or a list of pair of
+                string/string-sequences/int-sequence (see details in encode_plus)
+            add_special_tokens: if set to ``True``, the sequences will be encoded with the special tokens relative
+                to their model.
+            max_length: if set to a number, will limit the total sequence returned so that it has a maximum length.
+                If there are overflowing tokens, those will be added to the returned dictionary`
+            stride: if set to a number along with max_length, the overflowing tokens returned will contain some tokens
+                from the main sequence returned. The value of this argument defines the number of additional tokens.
+            truncation_strategy: string selected in the following options:
+                - 'longest_first' (default) Iteratively reduce the inputs sequence until the input is under max_length
+                    starting from the longest one at each token (when there is a pair of input sequences)
+                - 'only_first': Only truncate the first sequence
+                - 'only_second': Only truncate the second sequence
+                - 'do_not_truncate': Does not truncate (raise an error if the input sequence is longer than max_length)
+            return_tensors: (optional) can be set to 'tf' or 'pt' to return respectively TensorFlow tf.constant
+                or PyTorch torch.Tensor instead of a list of python integers.
+            **kwargs: passed to the `self.tokenize()` method
+        """
+        batch_outputs = {}
+        for ids_or_pair_ids in batch_text_or_text_pairs:
+            if isinstance(ids_or_pair_ids, (list, tuple)):
+                assert len(ids_or_pair_ids) == 2
+                ids, pair_ids = ids_or_pair_ids
+            else:
+                ids, pair_ids = ids_or_pair_ids, None
+            outputs = self.encode_plus(ids, pair_ids, add_special_tokens=add_special_tokens, max_length=max_length,
+                                       stride=stride, truncation_strategy=truncation_strategy, return_tensors=None)
+
+            # Append the non-padded length to the output
+            if return_input_lengths:
+                outputs['input_len'] = len(outputs['input_ids'])
+
+            for key, value in outputs.items():
+                if key not in batch_outputs:
+                    batch_outputs[key] = []
+                batch_outputs[key].append(value)
+
+        # Compute longest sequence size
+        max_seq_len = max(map(len, batch_outputs['input_ids']))
+
+        if return_attention_masks:
+            # Allow the model to not give any special attention to padded input
+            batch_outputs['attention_mask'] = [[0] * len(v) for v in batch_outputs['input_ids']]
+
+        if return_tensors is not None:
+
+            # Do the tensor conversion in batch
+            for key, value in batch_outputs.items():
+
+                padded_value = value
+                if key != 'input_len':
+                    # Padding handle
+                    padded_value = [v + [self.pad_token_id if key == 'input_ids' else 1] * (max_seq_len - len(v)) for v in padded_value]
+
+                if return_tensors == 'tf' and is_tf_available():
+                    batch_outputs[key] = tf.constant(padded_value)
+                elif return_tensors == 'pt' and is_torch_available():
+                    batch_outputs[key] = torch.tensor(padded_value)
+                elif return_tensors is not None:
+                    logger.warning("Unable to convert output to tensors format {}, PyTorch or TensorFlow is not available.".format(return_tensors))
+
+        # encoder_attention_mask requires 1 for real token, 0 for padding, just invert value
+        if return_attention_masks:
+            if is_tf_available():
+                batch_outputs['attention_mask'] = tf.abs(batch_outputs['attention_mask'] - 1)
+            else:
+                batch_outputs['attention_mask'] = torch.abs(batch_outputs['attention_mask'] - 1)
+
+        return batch_outputs
+
     def prepare_for_model(self, ids, pair_ids=None, max_length=None, add_special_tokens=True, stride=0,
                           truncation_strategy='longest_first',
                           pad_to_max_length=False,