Merge branch 'huggingface/master'

examples/run_multiple_choice.py

@@ -29,7 +29,12 @@ import torch
 from torch.utils.data import (DataLoader, RandomSampler, SequentialSampler,
                               TensorDataset)
 from torch.utils.data.distributed import DistributedSampler
-from tensorboardX import SummaryWriter
+
+try:
+    from torch.utils.tensorboard import SummaryWriter
+except:
+    from tensorboardX import SummaryWriter
+
 from tqdm import tqdm, trange
 
 from transformers import (WEIGHTS_NAME, BertConfig,
@@ -293,7 +298,7 @@ def load_and_cache_examples(args, task, tokenizer, evaluate=False, test=False):
         list(filter(None, args.model_name_or_path.split('/'))).pop(),
         str(args.max_seq_length),
         str(task)))
-    if os.path.exists(cached_features_file):
+    if os.path.exists(cached_features_file) and not args.overwrite_cache:
         logger.info("Loading features from cached file %s", cached_features_file)
         features = torch.load(cached_features_file)
     else:
@@ -306,14 +311,14 @@ def load_and_cache_examples(args, task, tokenizer, evaluate=False, test=False):
         else:
             examples = processor.get_train_examples(args.data_dir)
         logger.info("Training number: %s", str(len(examples)))
-        features = convert_examples_to_features(examples, label_list, args.max_seq_length, tokenizer,
-            cls_token_at_end=bool(args.model_type in ['xlnet']),            # xlnet has a cls token at the end
-            cls_token=tokenizer.cls_token,
-            sep_token=tokenizer.sep_token,
-            sep_token_extra=bool(args.model_type in ['roberta']),
-            cls_token_segment_id=2 if args.model_type in ['xlnet'] else 0,
+        features = convert_examples_to_features(
+            examples,
+            label_list,
+            args.max_seq_length,
+            tokenizer,
             pad_on_left=bool(args.model_type in ['xlnet']),                 # pad on the left for xlnet
-            pad_token_segment_id=4 if args.model_type in ['xlnet'] else 0)
+            pad_token_segment_id=4 if args.model_type in ['xlnet'] else 0
+        )
         if args.local_rank in [-1, 0]:
             logger.info("Saving features into cached file %s", cached_features_file)
             torch.save(features, cached_features_file)
@@ -362,7 +367,7 @@ def main():
                         help="Whether to run eval on the dev set.")
     parser.add_argument("--do_test", action='store_true', help='Whether to run test on the test set')
     parser.add_argument("--evaluate_during_training", action='store_true',
-                        help="Rul evaluation during training at each logging step.")
+                        help="Run evaluation during training at each logging step.")
     parser.add_argument("--do_lower_case", action='store_true',
                         help="Set this flag if you are using an uncased model.")
 
@@ -512,9 +517,11 @@ def main():
         logger.info("Evaluate the following checkpoints: %s", checkpoints)
         for checkpoint in checkpoints:
             global_step = checkpoint.split('-')[-1] if len(checkpoints) > 1 else ""
+            prefix = checkpoint.split('/')[-1] if checkpoint.find('checkpoint') != -1 else ""
+            
             model = model_class.from_pretrained(checkpoint)
             model.to(args.device)
-            result = evaluate(args, model, tokenizer, prefix=global_step)
+            result = evaluate(args, model, tokenizer, prefix=prefix)
             result = dict((k + '_{}'.format(global_step), v) for k, v in result.items())
             results.update(result)
 
@@ -528,9 +535,11 @@ def main():
         logger.info("Evaluate the following checkpoints: %s", checkpoints)
         for checkpoint in checkpoints:
             global_step = checkpoint.split('-')[-1] if len(checkpoints) > 1 else ""
+            prefix = checkpoint.split('/')[-1] if checkpoint.find('checkpoint') != -1 else ""
+            
             model = model_class.from_pretrained(checkpoint)
             model.to(args.device)
-            result = evaluate(args, model, tokenizer, prefix=global_step, test=True)
+            result = evaluate(args, model, tokenizer, prefix=prefix, test=True)
             result = dict((k + '_{}'.format(global_step), v) for k, v in result.items())
             results.update(result)
     if best_steps:

examples/run_squad.py

@@ -13,7 +13,7 @@
 # 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.
-""" Finetuning the library models for question-answering on SQuAD (Bert, XLM, XLNet)."""
+""" Finetuning the library models for question-answering on SQuAD (DistilBERT, Bert, XLM, XLNet)."""
 
 from __future__ import absolute_import, division, print_function
 
@@ -28,9 +28,13 @@ import torch
 from torch.utils.data import (DataLoader, RandomSampler, SequentialSampler,
                               TensorDataset)
 from torch.utils.data.distributed import DistributedSampler
-from tqdm import tqdm, trange
 
-from tensorboardX import SummaryWriter
+try:
+    from torch.utils.tensorboard import SummaryWriter
+except:
+    from tensorboardX import SummaryWriter
+
+from tqdm import tqdm, trange
 
 from transformers import (WEIGHTS_NAME, BertConfig,
                                   BertForQuestionAnswering, BertTokenizer,
@@ -135,9 +139,10 @@ def train(args, train_dataset, model, tokenizer):
             batch = tuple(t.to(args.device) for t in batch)
             inputs = {'input_ids':       batch[0],
                       'attention_mask':  batch[1],
-                      'token_type_ids':  None if args.model_type == 'xlm' else batch[2],  
                       'start_positions': batch[3],
                       'end_positions':   batch[4]}
+            if args.model_type != 'distilbert':
+                inputs['token_type_ids'] = None if args.model_type == 'xlm' else batch[2]
             if args.model_type in ['xlnet', 'xlm']:
                 inputs.update({'cls_index': batch[5],
                                'p_mask':       batch[6]})
@@ -218,9 +223,10 @@ def evaluate(args, model, tokenizer, prefix=""):
         batch = tuple(t.to(args.device) for t in batch)
         with torch.no_grad():
             inputs = {'input_ids':      batch[0],
-                      'attention_mask': batch[1],
-                      'token_type_ids': None if args.model_type == 'xlm' else batch[2]  # XLM don't use segment_ids
+                      'attention_mask': batch[1]
                       }
+            if args.model_type != 'distilbert':
+                inputs['token_type_ids'] = None if args.model_type == 'xlm' else batch[2]  # XLM don't use segment_ids
             example_indices = batch[3]
             if args.model_type in ['xlnet', 'xlm']:
                 inputs.update({'cls_index': batch[4],
@@ -475,6 +481,16 @@ def main():
 
     logger.info("Training/evaluation parameters %s", args)
 
+    # Before we do anything with models, we want to ensure that we get fp16 execution of torch.einsum if args.fp16 is set.
+    # Otherwise it'll default to "promote" mode, and we'll get fp32 operations. Note that running `--fp16_opt_level="O2"` will
+    # remove the need for this code, but it is still valid.
+    if args.fp16:
+        try:
+            import apex
+            apex.amp.register_half_function(torch, 'einsum')
+        except ImportError:
+            raise ImportError("Please install apex from https://www.github.com/nvidia/apex to use fp16 training.")
+
     # Training
     if args.do_train:
         train_dataset = load_and_cache_examples(args, tokenizer, evaluate=False, output_examples=False)

examples/run_tf_glue.py

+import os
 import tensorflow as tf
 import tensorflow_datasets
 from transformers import BertTokenizer, TFBertForSequenceClassification, glue_convert_examples_to_features, BertForSequenceClassification
 
-# Load dataset, tokenizer, model from pretrained model/vocabulary
+# script parameters
+BATCH_SIZE = 32
+EVAL_BATCH_SIZE = BATCH_SIZE * 2
+USE_XLA = False
+USE_AMP = False
+
+tf.config.optimizer.set_jit(USE_XLA)
+tf.config.optimizer.set_experimental_options({"auto_mixed_precision": USE_AMP})
+
+# Load tokenizer and model from pretrained model/vocabulary
 tokenizer = BertTokenizer.from_pretrained('bert-base-cased')
 model = TFBertForSequenceClassification.from_pretrained('bert-base-cased')
-data = tensorflow_datasets.load('glue/mrpc')
+
+# Load dataset via TensorFlow Datasets
+data, info = tensorflow_datasets.load('glue/mrpc', with_info=True)
+train_examples = info.splits['train'].num_examples
+valid_examples = info.splits['validation'].num_examples
 
 # Prepare dataset for GLUE as a tf.data.Dataset instance
 train_dataset = glue_convert_examples_to_features(data['train'], tokenizer, 128, 'mrpc')
 valid_dataset = glue_convert_examples_to_features(data['validation'], tokenizer, 128, 'mrpc')
-train_dataset = train_dataset.shuffle(100).batch(32).repeat(2)
-valid_dataset = valid_dataset.batch(64)
+train_dataset = train_dataset.shuffle(128).batch(BATCH_SIZE).repeat(-1)
+valid_dataset = valid_dataset.batch(EVAL_BATCH_SIZE)
 
 # Prepare training: Compile tf.keras model with optimizer, loss and learning rate schedule 
-optimizer = tf.keras.optimizers.Adam(learning_rate=3e-5, epsilon=1e-08, clipnorm=1.0)
+opt = tf.keras.optimizers.Adam(learning_rate=3e-5, epsilon=1e-08)
+if USE_AMP:
+    # loss scaling is currently required when using mixed precision
+    opt = tf.keras.mixed_precision.experimental.LossScaleOptimizer(opt, 'dynamic')
 loss = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True)
 metric = tf.keras.metrics.SparseCategoricalAccuracy('accuracy')
-model.compile(optimizer=optimizer, loss=loss, metrics=[metric])
+model.compile(optimizer=opt, loss=loss, metrics=[metric])
 
 # Train and evaluate using tf.keras.Model.fit()
-history = model.fit(train_dataset, epochs=2, steps_per_epoch=115,
-                    validation_data=valid_dataset, validation_steps=7)
+train_steps = train_examples//BATCH_SIZE
+valid_steps = valid_examples//EVAL_BATCH_SIZE
 
-# Load the TensorFlow model in PyTorch for inspection
+history = model.fit(train_dataset, epochs=2, steps_per_epoch=train_steps,
+                    validation_data=valid_dataset, validation_steps=valid_steps)
+
+# Save TF2 model
+os.makedirs('./save/', exist_ok=True)
 model.save_pretrained('./save/')
+
+# Load the TensorFlow model in PyTorch for inspection
 pytorch_model = BertForSequenceClassification.from_pretrained('./save/', from_tf=True)
 
 # Quickly test a few predictions - MRPC is a paraphrasing task, let's see if our model learned the task
-sentence_0 = "This research was consistent with his findings."
-sentence_1 = "His findings were compatible with this research."
-sentence_2 = "His findings were not compatible with this research."
+sentence_0 = 'This research was consistent with his findings.'
+sentence_1 = 'His findings were compatible with this research.'
+sentence_2 = 'His findings were not compatible with this research.'
 inputs_1 = tokenizer.encode_plus(sentence_0, sentence_1, add_special_tokens=True, return_tensors='pt')
 inputs_2 = tokenizer.encode_plus(sentence_0, sentence_2, add_special_tokens=True, return_tensors='pt')
 
 pred_1 = pytorch_model(**inputs_1)[0].argmax().item()
 pred_2 = pytorch_model(**inputs_2)[0].argmax().item()
-print("sentence_1 is", "a paraphrase" if pred_1 else "not a paraphrase", "of sentence_0")
-print("sentence_2 is", "a paraphrase" if pred_2 else "not a paraphrase", "of sentence_0")
+print('sentence_1 is', 'a paraphrase' if pred_1 else 'not a paraphrase', 'of sentence_0')
+print('sentence_2 is', 'a paraphrase' if pred_2 else 'not a paraphrase', 'of sentence_0')

examples/utils_multiple_choice.py

@@ -13,7 +13,7 @@
 # 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.
-""" BERT multiple choice fine-tuning: utilities to work with multiple choice tasks of reading comprehension  """
+""" Multiple choice fine-tuning: utilities to work with multiple choice tasks of reading comprehension  """
 
 from __future__ import absolute_import, division, print_function
 
@@ -26,6 +26,8 @@ import json
 import csv
 import glob
 import tqdm
+from typing import List
+from transformers import PreTrainedTokenizer
 
 
 logger = logging.getLogger(__name__)
@@ -40,7 +42,7 @@ class InputExample(object):
         Args:
             example_id: Unique id for the example.
             contexts: list of str. The untokenized text of the first sequence (context of corresponding question).
-            question: string. The untokenized text of the second sequence (qustion).
+            question: string. The untokenized text of the second sequence (question).
             endings: list of str. multiple choice's options. Its length must be equal to contexts' length.
             label: (Optional) string. The label of the example. This should be
             specified for train and dev examples, but not for test examples.
@@ -66,7 +68,7 @@ class InputFeatures(object):
                 'input_mask': input_mask,
                 'segment_ids': segment_ids
             }
-            for _, input_ids, input_mask, segment_ids in choices_features
+            for input_ids, input_mask, segment_ids in choices_features
         ]
         self.label = label
 
@@ -192,7 +194,7 @@ class SwagProcessor(DataProcessor):
             return lines
 
 
-    def _create_examples(self, lines, type):
+    def _create_examples(self, lines: List[List[str]], type: str):
         """Creates examples for the training and dev sets."""
         if type == "train" and lines[0][-1] != 'label':
             raise ValueError(
@@ -300,24 +302,18 @@ class ArcProcessor(DataProcessor):
         return examples
 
 
-def convert_examples_to_features(examples, label_list, max_seq_length,
-                                 tokenizer,
-                                 cls_token_at_end=False,
-                                 cls_token='[CLS]',
-                                 cls_token_segment_id=1,
-                                 sep_token='[SEP]',
-                                 sequence_a_segment_id=0,
-                                 sequence_b_segment_id=1,
-                                 sep_token_extra=False,
+def convert_examples_to_features(
+    examples: List[InputExample],
+    label_list: List[str],
+    max_length: int,
+    tokenizer: PreTrainedTokenizer,
     pad_token_segment_id=0,
     pad_on_left=False,
     pad_token=0,
-                                 mask_padding_with_zero=True):
-    """ Loads a data file into a list of `InputBatch`s
-        `cls_token_at_end` define the location of the CLS token:
-            - False (Default, BERT/XLM pattern): [CLS] + A + [SEP] + B + [SEP]
-            - True (XLNet/GPT pattern): A + [SEP] + B + [SEP] + [CLS]
-        `cls_token_segment_id` define the segment id associated to the CLS token (0 for BERT, 2 for XLNet)
+    mask_padding_with_zero=True,
+) -> List[InputFeatures]:
+    """
+    Loads a data file into a list of `InputFeatures`
     """
 
     label_map = {label : i for i, label in enumerate(label_list)}
@@ -328,125 +324,70 @@ def convert_examples_to_features(examples, label_list, max_seq_length,
             logger.info("Writing example %d of %d" % (ex_index, len(examples)))
         choices_features = []
         for ending_idx, (context, ending) in enumerate(zip(example.contexts, example.endings)):
-            tokens_a = tokenizer.tokenize(context)
-            tokens_b = None
+            text_a = context
             if example.question.find("_") != -1:
-                #this is for cloze question
-                tokens_b = tokenizer.tokenize(example.question.replace("_", ending))
-            else:
-                tokens_b = tokenizer.tokenize(example.question + " " + ending)
-                # you can add seq token between quesiotn and ending. This does not make too much difference.
-                # tokens_b = tokenizer.tokenize(example.question)
-                # tokens_b += [sep_token]
-                # if sep_token_extra:
-                #     tokens_b += [sep_token]
-                # tokens_b += tokenizer.tokenize(ending)
-
-            special_tokens_count = 4 if sep_token_extra else 3
-            _truncate_seq_pair(tokens_a, tokens_b, max_seq_length - special_tokens_count)
-
-            # The convention in BERT is:
-            # (a) For sequence pairs:
-            #  tokens:   [CLS] is this jack ##son ##ville ? [SEP] no it is not . [SEP]
-            #  type_ids:   0   0  0    0    0     0       0   0   1  1  1  1   1   1
-            # (b) For single sequences:
-            #  tokens:   [CLS] the dog is hairy . [SEP]
-            #  type_ids:   0   0   0   0  0     0   0
-            #
-            # Where "type_ids" are used to indicate whether this is the first
-            # sequence or the second sequence. The embedding vectors for `type=0` and
-            # `type=1` were learned during pre-training and are added to the wordpiece
-            # embedding vector (and position vector). This is not *strictly* necessary
-            # since the [SEP] token unambiguously separates the sequences, but it makes
-            # it easier for the model to learn the concept of sequences.
-            #
-            # For classification tasks, the first vector (corresponding to [CLS]) is
-            # used as as the "sentence vector". Note that this only makes sense because
-            # the entire model is fine-tuned.
-            tokens = tokens_a + [sep_token]
-            if sep_token_extra:
-                # roberta uses an extra separator b/w pairs of sentences
-                tokens += [sep_token]
-
-            segment_ids = [sequence_a_segment_id] * len(tokens)
-
-            if tokens_b:
-                tokens += tokens_b + [sep_token]
-                segment_ids += [sequence_b_segment_id] * (len(tokens_b) + 1)
-
-            if cls_token_at_end:
-                tokens = tokens + [cls_token]
-                segment_ids = segment_ids + [cls_token_segment_id]
+                # this is for cloze question
+                text_b = example.question.replace("_", ending)
             else:
-                tokens = [cls_token] + tokens
-                segment_ids = [cls_token_segment_id] + segment_ids
+                text_b = example.question + " " + ending
+
+            inputs = tokenizer.encode_plus(
+                text_a,
+                text_b,
+                add_special_tokens=True,
+                max_length=max_length,
+            )
+            if 'num_truncated_tokens' in inputs and inputs['num_truncated_tokens'] > 0:
+                logger.info('Attention! you are cropping tokens (swag task is ok). '
+                        'If you are training ARC and RACE and you are poping question + options,'
+                        'you need to try to use a bigger max seq length!')
 
-            input_ids = tokenizer.convert_tokens_to_ids(tokens)
+            input_ids, token_type_ids = inputs["input_ids"], inputs["token_type_ids"]
 
             # The mask has 1 for real tokens and 0 for padding tokens. Only real
             # tokens are attended to.
-            input_mask = [1 if mask_padding_with_zero else 0] * len(input_ids)
+            attention_mask = [1 if mask_padding_with_zero else 0] * len(input_ids)
 
             # Zero-pad up to the sequence length.
-            padding_length = max_seq_length - len(input_ids)
+            padding_length = max_length - len(input_ids)
             if pad_on_left:
                 input_ids = ([pad_token] * padding_length) + input_ids
-                input_mask = ([0 if mask_padding_with_zero else 1] * padding_length) + input_mask
-                segment_ids = ([pad_token_segment_id] * padding_length) + segment_ids
+                attention_mask = ([0 if mask_padding_with_zero else 1] * padding_length) + attention_mask
+                token_type_ids = ([pad_token_segment_id] * padding_length) + token_type_ids
             else:
                 input_ids = input_ids + ([pad_token] * padding_length)
-                input_mask = input_mask + ([0 if mask_padding_with_zero else 1] * padding_length)
-                segment_ids = segment_ids + ([pad_token_segment_id] * padding_length)
+                attention_mask = attention_mask + ([0 if mask_padding_with_zero else 1] * padding_length)
+                token_type_ids = token_type_ids + ([pad_token_segment_id] * padding_length)
+
+            assert len(input_ids) == max_length
+            assert len(attention_mask) == max_length
+            assert len(token_type_ids) == max_length
+            choices_features.append((input_ids, attention_mask, token_type_ids))
+
 
-            assert len(input_ids) == max_seq_length
-            assert len(input_mask) == max_seq_length
-            assert len(segment_ids) == max_seq_length
-            choices_features.append((tokens, input_ids, input_mask, segment_ids))
         label = label_map[example.label]
 
         if ex_index < 2:
             logger.info("*** Example ***")
             logger.info("race_id: {}".format(example.example_id))
-            for choice_idx, (tokens, input_ids, input_mask, segment_ids) in enumerate(choices_features):
+            for choice_idx, (input_ids, attention_mask, token_type_ids) in enumerate(choices_features):
                 logger.info("choice: {}".format(choice_idx))
-                logger.info("tokens: {}".format(' '.join(tokens)))
                 logger.info("input_ids: {}".format(' '.join(map(str, input_ids))))
-                logger.info("input_mask: {}".format(' '.join(map(str, input_mask))))
-                logger.info("segment_ids: {}".format(' '.join(map(str, segment_ids))))
+                logger.info("attention_mask: {}".format(' '.join(map(str, attention_mask))))
+                logger.info("token_type_ids: {}".format(' '.join(map(str, token_type_ids))))
                 logger.info("label: {}".format(label))
 
         features.append(
             InputFeatures(
-                example_id = example.example_id,
-                choices_features = choices_features,
-                label = label
+                example_id=example.example_id,
+                choices_features=choices_features,
+                label=label,
             )
         )
 
     return features
 
 
-def _truncate_seq_pair(tokens_a, tokens_b, max_length):
-    """Truncates a sequence pair in place to the maximum length."""
-
-    # This is a simple heuristic which will always truncate the longer sequence
-    # one token at a time. This makes more sense than truncating an equal percent
-    # of tokens from each, since if one sequence is very short then each token
-    # that's truncated likely contains more information than a longer sequence.
-
-    # However, since we'd better not to remove tokens of options and questions, you can choose to use a bigger
-    # length or only pop from context
-    while True:
-        total_length = len(tokens_a) + len(tokens_b)
-        if total_length <= max_length:
-            break
-        if len(tokens_a) > len(tokens_b):
-            tokens_a.pop()
-        else:
-            logger.info('Attention! you are removing from token_b (swag task is ok). '
-                        'If you are training ARC and RACE (you are poping question + options), '
-                        'you need to try to use a bigger max seq length!')
-            tokens_b.pop()
 
 
 processors = {
@@ -456,7 +397,7 @@ processors = {
 }
 
 
-GLUE_TASKS_NUM_LABELS = {
+MULTIPLE_CHOICE_TASKS_NUM_LABELS = {
     "race", 4,
     "swag", 4,
     "arc", 4

examples/utils_ner.py

+# coding=utf-8
+# Copyright 2018 The Google AI Language Team Authors and The HuggingFace Inc. team.
+# Copyright (c) 2018, NVIDIA CORPORATION.  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.
+""" Named entity recognition fine-tuning: utilities to work with CoNLL-2003 task. """
+
+from __future__ import absolute_import, division, print_function
+
+import logging
+import os
+from io import open
+
+logger = logging.getLogger(__name__)
+
+
+class InputExample(object):
+    """A single training/test example for token classification."""
+
+    def __init__(self, guid, words, labels):
+        """Constructs a InputExample.
+
+        Args:
+            guid: Unique id for the example.
+            words: list. The words of the sequence.
+            labels: (Optional) list. The labels for each word of the sequence. This should be
+            specified for train and dev examples, but not for test examples.
+        """
+        self.guid = guid
+        self.words = words
+        self.labels = labels
+
+
+class InputFeatures(object):
+    """A single set of features of data."""
+
+    def __init__(self, input_ids, input_mask, segment_ids, label_ids):
+        self.input_ids = input_ids
+        self.input_mask = input_mask
+        self.segment_ids = segment_ids
+        self.label_ids = label_ids
+
+
+def read_examples_from_file(data_dir, mode):
+    file_path = os.path.join(data_dir, "{}.txt".format(mode))
+    guid_index = 1
+    examples = []
+    with open(file_path, encoding="utf-8") as f:
+        words = []
+        labels = []
+        for line in f:
+            if line.startswith("-DOCSTART-") or line == "" or line == "\n":
+                if words:
+                    examples.append(InputExample(guid="{}-{}".format(mode, guid_index),
+                                                 words=words,
+                                                 labels=labels))
+                    guid_index += 1
+                    words = []
+                    labels = []
+            else:
+                splits = line.split(" ")
+                words.append(splits[0])
+                if len(splits) > 1:
+                    labels.append(splits[-1].replace("\n", ""))
+                else:
+                    # Examples could have no label for mode = "test"
+                    labels.append("O")
+        if words:
+            examples.append(InputExample(guid="%s-%d".format(mode, guid_index),
+                                         words=words,
+                                         labels=labels))
+    return examples
+
+
+def convert_examples_to_features(examples,
+                                 label_list,
+                                 max_seq_length,
+                                 tokenizer,
+                                 cls_token_at_end=False,
+                                 cls_token="[CLS]",
+                                 cls_token_segment_id=1,
+                                 sep_token="[SEP]",
+                                 sep_token_extra=False,
+                                 pad_on_left=False,
+                                 pad_token=0,
+                                 pad_token_segment_id=0,
+                                 pad_token_label_id=-1,
+                                 sequence_a_segment_id=0,
+                                 mask_padding_with_zero=True):
+    """ Loads a data file into a list of `InputBatch`s
+        `cls_token_at_end` define the location of the CLS token:
+            - False (Default, BERT/XLM pattern): [CLS] + A + [SEP] + B + [SEP]
+            - True (XLNet/GPT pattern): A + [SEP] + B + [SEP] + [CLS]
+        `cls_token_segment_id` define the segment id associated to the CLS token (0 for BERT, 2 for XLNet)
+    """
+
+    label_map = {label: i for i, label in enumerate(label_list)}
+
+    features = []
+    for (ex_index, example) in enumerate(examples):
+        if ex_index % 10000 == 0:
+            logger.info("Writing example %d of %d", ex_index, len(examples))
+
+        tokens = []
+        label_ids = []
+        for word, label in zip(example.words, example.labels):
+            word_tokens = tokenizer.tokenize(word)
+            tokens.extend(word_tokens)
+            # Use the real label id for the first token of the word, and padding ids for the remaining tokens
+            label_ids.extend([label_map[label]] + [pad_token_label_id] * (len(word_tokens) - 1))
+
+        # Account for [CLS] and [SEP] with "- 2" and with "- 3" for RoBERTa.
+        special_tokens_count = 3 if sep_token_extra else 2
+        if len(tokens) > max_seq_length - special_tokens_count:
+            tokens = tokens[:(max_seq_length - special_tokens_count)]
+            label_ids = label_ids[:(max_seq_length - special_tokens_count)]
+
+        # The convention in BERT is:
+        # (a) For sequence pairs:
+        #  tokens:   [CLS] is this jack ##son ##ville ? [SEP] no it is not . [SEP]
+        #  type_ids:   0   0  0    0    0     0       0   0   1  1  1  1   1   1
+        # (b) For single sequences:
+        #  tokens:   [CLS] the dog is hairy . [SEP]
+        #  type_ids:   0   0   0   0  0     0   0
+        #
+        # Where "type_ids" are used to indicate whether this is the first
+        # sequence or the second sequence. The embedding vectors for `type=0` and
+        # `type=1` were learned during pre-training and are added to the wordpiece
+        # embedding vector (and position vector). This is not *strictly* necessary
+        # since the [SEP] token unambiguously separates the sequences, but it makes
+        # it easier for the model to learn the concept of sequences.
+        #
+        # For classification tasks, the first vector (corresponding to [CLS]) is
+        # used as as the "sentence vector". Note that this only makes sense because
+        # the entire model is fine-tuned.
+        tokens += [sep_token]
+        label_ids += [pad_token_label_id]
+        if sep_token_extra:
+            # roberta uses an extra separator b/w pairs of sentences
+            tokens += [sep_token]
+            label_ids += [pad_token_label_id]
+        segment_ids = [sequence_a_segment_id] * len(tokens)
+
+        if cls_token_at_end:
+            tokens += [cls_token]
+            label_ids += [pad_token_label_id]
+            segment_ids += [cls_token_segment_id]
+        else:
+            tokens = [cls_token] + tokens
+            label_ids = [pad_token_label_id] + label_ids
+            segment_ids = [cls_token_segment_id] + segment_ids
+
+        input_ids = tokenizer.convert_tokens_to_ids(tokens)
+
+        # The mask has 1 for real tokens and 0 for padding tokens. Only real
+        # tokens are attended to.
+        input_mask = [1 if mask_padding_with_zero else 0] * len(input_ids)
+
+        # Zero-pad up to the sequence length.
+        padding_length = max_seq_length - len(input_ids)
+        if pad_on_left:
+            input_ids = ([pad_token] * padding_length) + input_ids
+            input_mask = ([0 if mask_padding_with_zero else 1] * padding_length) + input_mask
+            segment_ids = ([pad_token_segment_id] * padding_length) + segment_ids
+            label_ids = ([pad_token_label_id] * padding_length) + label_ids
+        else:
+            input_ids += ([pad_token] * padding_length)
+            input_mask += ([0 if mask_padding_with_zero else 1] * padding_length)
+            segment_ids += ([pad_token_segment_id] * padding_length)
+            label_ids += ([pad_token_label_id] * padding_length)
+
+        assert len(input_ids) == max_seq_length
+        assert len(input_mask) == max_seq_length
+        assert len(segment_ids) == max_seq_length
+        assert len(label_ids) == max_seq_length
+
+        if ex_index < 5:
+            logger.info("*** Example ***")
+            logger.info("guid: %s", example.guid)
+            logger.info("tokens: %s", " ".join([str(x) for x in tokens]))
+            logger.info("input_ids: %s", " ".join([str(x) for x in input_ids]))
+            logger.info("input_mask: %s", " ".join([str(x) for x in input_mask]))
+            logger.info("segment_ids: %s", " ".join([str(x) for x in segment_ids]))
+            logger.info("label_ids: %s", " ".join([str(x) for x in label_ids]))
+
+        features.append(
+                InputFeatures(input_ids=input_ids,
+                              input_mask=input_mask,
+                              segment_ids=segment_ids,
+                              label_ids=label_ids))
+    return features
+
+
+def get_labels(path):
+    if path:
+        with open(path, "r") as f:
+            labels = f.read().splitlines()
+        if "O" not in labels:
+            labels = ["O"] + labels
+        return labels
+    else:
+        return ["O", "B-MISC", "I-MISC",  "B-PER", "I-PER", "B-ORG", "I-ORG", "B-LOC", "I-LOC"]

requirements-dev.txt

+absl-py==0.8.0
+astor==0.8.0
+atomicwrites==1.3.0
+attrs==19.2.0
+boto3==1.9.243
+botocore==1.12.243
+certifi==2019.9.11
+chardet==3.0.4
+Click==7.0
+docutils==0.15.2
+gast==0.2.2
+google-pasta==0.1.7
+grpcio==1.24.1
+h5py==2.10.0
+idna==2.8
+importlib-metadata==0.23
+jmespath==0.9.4
+joblib==0.14.0
+Keras-Applications==1.0.8
+Keras-Preprocessing==1.1.0
+Markdown==3.1.1
+more-itertools==7.2.0
+numpy==1.17.2
+opt-einsum==3.1.0
+packaging==19.2
+pluggy==0.13.0
+protobuf==3.10.0
+py==1.8.0
+pyparsing==2.4.2
+pytest==5.2.1
+python-dateutil==2.8.0
+regex==2019.8.19
+requests==2.22.0
+s3transfer==0.2.1
+sacremoses==0.0.35
+sentencepiece==0.1.83
+six==1.12.0
+tensorboard==2.0.0
+tensorflow==2.0.0
+tensorflow-estimator==2.0.0
+termcolor==1.1.0
+torch==1.2.0
+tqdm==4.36.1
+urllib3==1.25.6
+wcwidth==0.1.7
+Werkzeug==0.16.0
+wrapt==1.11.2
+zipp==0.6.0

setup.py

@@ -3,7 +3,7 @@ Simple check list from AllenNLP repo: https://github.com/allenai/allennlp/blob/m
 
 To create the package for pypi.
 
-1. Change the version in __init__.py and setup.py.
+1. Change the version in __init__.py, setup.py as well as docs/source/conf.py.
 
 2. Commit these changes with the message: "Release: VERSION"
 
@@ -38,7 +38,7 @@ from setuptools import find_packages, setup
 
 setup(
     name="transformers",
-    version="2.0.0",
+    version="2.1.1",
     author="Thomas Wolf, Lysandre Debut, Victor Sanh, Julien Chaumond, Google AI Language Team Authors, Open AI team Authors, Facebook AI Authors, Carnegie Mellon University Authors",
     author_email="thomas@huggingface.co",
     description="State-of-the-art Natural Language Processing for TensorFlow 2.0 and PyTorch",

transformers/configuration_bert.py

@@ -40,6 +40,8 @@ BERT_PRETRAINED_CONFIG_ARCHIVE_MAP = {
     'bert-large-uncased-whole-word-masking-finetuned-squad': "https://s3.amazonaws.com/models.huggingface.co/bert/bert-large-uncased-whole-word-masking-finetuned-squad-config.json",
     'bert-large-cased-whole-word-masking-finetuned-squad': "https://s3.amazonaws.com/models.huggingface.co/bert/bert-large-cased-whole-word-masking-finetuned-squad-config.json",
     'bert-base-cased-finetuned-mrpc': "https://s3.amazonaws.com/models.huggingface.co/bert/bert-base-cased-finetuned-mrpc-config.json",
+    'bert-base-german-dbmdz-cased': "https://s3.amazonaws.com/models.huggingface.co/bert/bert-base-german-dbmdz-cased-config.json",
+    'bert-base-german-dbmdz-uncased': "https://s3.amazonaws.com/models.huggingface.co/bert/bert-base-german-dbmdz-uncased-config.json",
 }