[API] Make BERT a hybrid block (#877)

* change BaseTransformerEncoder forward

* fix

* merge _forward to hybrid_forward

* if arange_like is not avaiable

* fix

* make transformer encoder a true hybridblock

* fix lint

* fix bug

* use zeroslike and infer_range to avoid backprop problem

* update test

* Revert "update test"

This reverts commit d024a0c.

* more printing

* fix test case

* make bert hybrid

* add legacy model

* fix lint

* fix lint

* revert change for default parameter

* fix arange dtype

* fix lint

* revert mokey patch in NMT interface

* fix dtype

* fix bug in arange

* remove legacy model

* also update bert scripts

* fix lint

* fix typo

* remove hybridbert test

scripts/bert/export/export.py → scripts/bert/export.py

@@ -43,8 +43,9 @@
 
 import mxnet as mx
 import gluonnlp as nlp
-from hybrid_bert import get_hybrid_model
-from hybrid_bert import HybridBERTClassifier, HybridBERTRegression, HybridBERTForQA
+from gluonnlp.model import get_model
+from model.classification import BERTClassifier, BERTRegression
+from model.qa import BertForQA
 
 parser = argparse.ArgumentParser(description='Export hybrid BERT base model.')
 
@@ -126,35 +127,35 @@
 seq_length = args.seq_length
 
 if args.task == 'classification':
-    bert, _ = get_hybrid_model(
+    bert, _ = get_model(
         name=args.model_name,
         dataset_name=args.dataset_name,
         pretrained=False,
         use_pooler=True,
         use_decoder=False,
         use_classifier=False,
         seq_length=args.seq_length)
-    net = HybridBERTClassifier(bert, num_classes=2, dropout=args.dropout)
+    net = BERTClassifier(bert, num_classes=2, dropout=args.dropout)
 elif args.task == 'regression':
-    bert, _ = get_hybrid_model(
+    bert, _ = get_model(
         name=args.model_name,
         dataset_name=args.dataset_name,
         pretrained=False,
         use_pooler=True,
         use_decoder=False,
         use_classifier=False,
         seq_length=args.seq_length)
-    net = HybridBERTRegression(bert, dropout=args.dropout)
+    net = BERTRegression(bert, dropout=args.dropout)
 elif args.task == 'question_answering':
-    bert, _ = get_hybrid_model(
+    bert, _ = get_model(
         name=args.model_name,
         dataset_name=args.dataset_name,
         pretrained=False,
         use_pooler=False,
         use_decoder=False,
         use_classifier=False,
         seq_length=args.seq_length)
-    net = HybridBERTForQA(bert)
+    net = BertForQA(bert)
 else:
     raise ValueError('unknown task: %s'%args.task)
 
@@ -163,7 +164,7 @@
 else:
     net.initialize()
     warnings.warn('--model_parameters is not provided. The parameter checkpoint (.params) '
-                  'file will be created based on default parameter intialization.')
+                  'file will be created based on default parameter initialization.')
 
 net.hybridize(static_alloc=True, static_shape=True)