Merge branch 'master' into fix_top_k_top_p_filtering

examples/distillation/scripts/extract_for_distil.py → examples/distillation/scripts/extract_distilbert.py

@@ -14,14 +14,15 @@
 # limitations under the License.
 """
 Preprocessing script before training DistilBERT.
+Specific to BERT -> DistilBERT.
 """
-from pytorch_transformers import BertForMaskedLM, RobertaForMaskedLM
+from transformers import BertForMaskedLM, RobertaForMaskedLM
 import torch
 import argparse
 
 if __name__ == '__main__':
     parser = argparse.ArgumentParser(description="Extraction some layers of the full BertForMaskedLM or RObertaForMaskedLM for Transfer Learned Distillation")
-    parser.add_argument("--model_type", default="bert", choices=["bert", "roberta"])
+    parser.add_argument("--model_type", default="bert", choices=["bert"])
     parser.add_argument("--model_name", default='bert-base-uncased', type=str)
     parser.add_argument("--dump_checkpoint", default='serialization_dir/tf_bert-base-uncased_0247911.pth', type=str)
     parser.add_argument("--vocab_transform", action='store_true')
@@ -31,9 +32,8 @@ if __name__ == '__main__':
     if args.model_type == 'bert':
         model = BertForMaskedLM.from_pretrained(args.model_name)
         prefix = 'bert'
-    elif args.model_type == 'roberta':
-        model = RobertaForMaskedLM.from_pretrained(args.model_name)
-        prefix = 'roberta'
+    else:
+        raise ValueError(f'args.model_type should be "bert".')
 
     state_dict = model.state_dict()
     compressed_sd = {}
@@ -68,20 +68,12 @@ if __name__ == '__main__':
                 state_dict[f'{prefix}.encoder.layer.{teacher_idx}.output.LayerNorm.{w}']
         std_idx += 1
 
-    if args.model_type == 'bert':
-        compressed_sd[f'vocab_projector.weight'] = state_dict[f'cls.predictions.decoder.weight']
-        compressed_sd[f'vocab_projector.bias'] = state_dict[f'cls.predictions.bias']
-        if args.vocab_transform:
-            for w in ['weight', 'bias']:
-                compressed_sd[f'vocab_transform.{w}'] = state_dict[f'cls.predictions.transform.dense.{w}']
-                compressed_sd[f'vocab_layer_norm.{w}'] = state_dict[f'cls.predictions.transform.LayerNorm.{w}']
-    elif args.model_type == 'roberta':
-        compressed_sd[f'vocab_projector.weight'] = state_dict[f'lm_head.decoder.weight']
-        compressed_sd[f'vocab_projector.bias'] = state_dict[f'lm_head.bias']
-        if args.vocab_transform:
-            for w in ['weight', 'bias']:
-                compressed_sd[f'vocab_transform.{w}'] = state_dict[f'lm_head.dense.{w}']
-                compressed_sd[f'vocab_layer_norm.{w}'] = state_dict[f'lm_head.layer_norm.{w}']
+    compressed_sd[f'vocab_projector.weight'] = state_dict[f'cls.predictions.decoder.weight']
+    compressed_sd[f'vocab_projector.bias'] = state_dict[f'cls.predictions.bias']
+    if args.vocab_transform:
+        for w in ['weight', 'bias']:
+            compressed_sd[f'vocab_transform.{w}'] = state_dict[f'cls.predictions.transform.dense.{w}']
+            compressed_sd[f'vocab_layer_norm.{w}'] = state_dict[f'cls.predictions.transform.LayerNorm.{w}']
 
     print(f'N layers selected for distillation: {std_idx}')
     print(f'Number of params transfered for distillation: {len(compressed_sd.keys())}')

examples/distillation/scripts/token_counts.py

@@ -13,7 +13,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 """
-Preprocessing script before training DistilBERT.
+Preprocessing script before training the distilled model.
 """
 from collections import Counter
 import argparse

examples/distillation/training_configs/distilbert-base-uncased.json

+{
+	"activation": "gelu",
+	"attention_dropout": 0.1,
+	"dim": 768,
+	"dropout": 0.1,
+	"hidden_dim": 3072,
+	"initializer_range": 0.02,
+	"max_position_embeddings": 512,
+	"n_heads": 12,
+	"n_layers": 6,
+	"sinusoidal_pos_embds": true,
+	"tie_weights_": true,
+	"vocab_size": 30522
+  }
+  
\ No newline at end of file

examples/distillation/training_configs/distilgpt2.json

+{
+	"initializer_range": 0.02,
+	"layer_norm_epsilon": 0.00001,
+	"n_ctx": 1024,
+	"n_embd": 768,
+	"n_head": 12,
+	"n_layer": 6,
+	"n_positions": 1024,
+	"vocab_size": 50257
+}
\ No newline at end of file

examples/requirements.txt

 tensorboardX
-scikit-learn
\ No newline at end of file
+tensorboard
+scikit-learn
+seqeval

examples/run_bertology.py

@@ -32,7 +32,7 @@ from torch.utils.data import DataLoader, SequentialSampler, TensorDataset, Subse
 from torch.utils.data.distributed import DistributedSampler
 from torch.nn import CrossEntropyLoss, MSELoss
 
-from pytorch_transformers import (WEIGHTS_NAME,
+from transformers import (WEIGHTS_NAME,
                                   BertConfig, BertForSequenceClassification, BertTokenizer,
                                   XLMConfig, XLMForSequenceClassification, XLMTokenizer,
                                   XLNetConfig, XLNetForSequenceClassification, XLNetTokenizer)

examples/run_multiple_choice.py

@@ -29,16 +29,21 @@ 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 pytorch_transformers import (WEIGHTS_NAME, BertConfig,
+from transformers import (WEIGHTS_NAME, BertConfig,
                                   BertForMultipleChoice, BertTokenizer,
                                   XLNetConfig, XLNetForMultipleChoice,
                                   XLNetTokenizer, RobertaConfig,
                                   RobertaForMultipleChoice, RobertaTokenizer)
 
-from pytorch_transformers import AdamW, WarmupLinearSchedule
+from transformers import AdamW, WarmupLinearSchedule
 
 from utils_multiple_choice import (convert_examples_to_features, processors)
 
@@ -141,7 +146,7 @@ def train(args, train_dataset, model, tokenizer):
                       'token_type_ids': batch[2] if args.model_type in ['bert', 'xlnet'] else None,  # XLM don't use segment_ids
                       'labels':         batch[3]}
             outputs = model(**inputs)
-            loss = outputs[0]  # model outputs are always tuple in pytorch-transformers (see doc)
+            loss = outputs[0]  # model outputs are always tuple in transformers (see doc)
 
             if args.n_gpu > 1:
                 loss = loss.mean() # mean() to average on multi-gpu parallel training
@@ -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.")
 
@@ -508,13 +513,15 @@ def main():
         checkpoints = [args.output_dir]
         if args.eval_all_checkpoints:
             checkpoints = list(os.path.dirname(c) for c in sorted(glob.glob(args.output_dir + '/**/' + WEIGHTS_NAME, recursive=True)))
-            logging.getLogger("pytorch_transformers.modeling_utils").setLevel(logging.WARN)  # Reduce logging
+            logging.getLogger("transformers.modeling_utils").setLevel(logging.WARN)  # Reduce logging
         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)
 
@@ -524,13 +531,15 @@ def main():
         checkpoints = [args.output_dir]
         # if args.eval_all_checkpoints: # can not use this to do test!!
         #     checkpoints = list(os.path.dirname(c) for c in sorted(glob.glob(args.output_dir + '/**/' + WEIGHTS_NAME, recursive=True)))
-        #     logging.getLogger("pytorch_transformers.modeling_utils").setLevel(logging.WARN)  # Reduce logging
+        #     logging.getLogger("transformers.modeling_utils").setLevel(logging.WARN)  # Reduce logging
         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/utils_squad.py

@@ -24,7 +24,7 @@ import math
 import collections
 from io import open
 
-from pytorch_transformers.tokenization_bert import BasicTokenizer, whitespace_tokenize
+from transformers.tokenization_bert import BasicTokenizer, whitespace_tokenize
 
 # Required by XLNet evaluation method to compute optimal threshold (see write_predictions_extended() method)
 from utils_squad_evaluate import find_all_best_thresh_v2, make_qid_to_has_ans, get_raw_scores