BigBird added for NER and classification

CHANGELOG.md

@@ -4,10 +4,20 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
-## [0.61.5] - 2021-05-18
+## [0.61.6] - 2021-05-28
+
+### Fixed
+
+- Fixed the onnx predict loop [whr778](https://github.com/whr778)
 
 ### Added
 
+- Added NER support for BigBird, Deberta, Deberta-v2, and xlm pretrained models [whr778](https://github.com/whr778)
+- Added BigBird for regular sequence classification (not multilabel) [@manueltonneau](https://github.com/manueltonneau)
+
+## [0.61.5] - 2021-05-18
+### Added
+
 - Fixed possible bug when using HF Datasets with Seq2SeqModel
 - Added `repo: simpletransformers` to W&B config
 
@@ -1504,7 +1514,9 @@ Model checkpoint is now saved for all epochs again.
 
 - This CHANGELOG file to hopefully serve as an evolving example of a standardized open source project CHANGELOG.
 
-[0.61.5]: https://github.com/ThilinaRajapakse/simpletransformers/compare/b49bf28...HEAD
+[0.61.6]: https://github.com/ThilinaRajapakse/simpletransformers/compare/281ff31...HEAD
+
+[0.61.5]: https://github.com/ThilinaRajapakse/simpletransformers/compare/b49bf28...281ff31
 
 [0.61.4]: https://github.com/ThilinaRajapakse/simpletransformers/compare/87eeb0e...b49bf28
 

setup.py

@@ -5,7 +5,7 @@
 
 setup(
     name="simpletransformers",
-    version="0.61.5",
+    version="0.61.6",
     author="Thilina Rajapakse",
     author_email="[email protected]",
     description="An easy-to-use wrapper library for the Transformers library.",

tests/test_classification.py

@@ -11,6 +11,7 @@
     "model_type, model_name",
     [
         ("bert", "bert-base-uncased"),
+        ("bigbird", "google/bigbird-roberta-base"),
         # ("longformer", "allenai/longformer-base-4096"),
         # ("electra", "google/electra-small-discriminator"),
         # ("mobilebert", "google/mobilebert-uncased"),
@@ -70,6 +71,7 @@ def test_binary_classification(model_type, model_name):
     [
         # ("bert", "bert-base-uncased"),
         # ("xlnet", "xlnet-base-cased"),
+        ("bigbird", "google/bigbird-roberta-base"),
         # ("xlm", "xlm-mlm-17-1280"),
         ("roberta", "roberta-base"),
         # ("distilbert", "distilbert-base-uncased"),
@@ -175,7 +177,22 @@ def test_multilabel_classification(model_type, model_name):
     )
 
 
-def test_sliding_window():
+@pytest.mark.parametrize(
+    "model_type, model_name",
+    [
+        # ("bert", "bert-base-uncased"),
+        # ("xlnet", "xlnet-base-cased"),
+        ("bigbird", "google/bigbird-roberta-base"),
+        # ("xlm", "xlm-mlm-17-1280"),
+        ("roberta", "roberta-base"),
+        # ("distilbert", "distilbert-base-uncased"),
+        # ("albert", "albert-base-v1"),
+        # ("camembert", "camembert-base"),
+        # ("xlmroberta", "xlm-roberta-base"),
+        # ("flaubert", "flaubert-base-cased"),
+    ],
+)
+def test_sliding_window(model_type, model_name):
     # Train and Evaluation data needs to be in a Pandas Dataframe of two columns.
     # The first column is the text with type str, and the second column is the
     # label with type int.
@@ -193,8 +210,8 @@ def test_sliding_window():
 
     # Create a ClassificationModel
     model = ClassificationModel(
-        "distilbert",
-        "distilbert-base-uncased",
+        model_type,
+        model_name,
         use_cuda=False,
         args={
             "no_save": True,

tests/test_named_entity_recognition.py

@@ -8,10 +8,11 @@
     "model_type, model_name",
     [
         ("bert", "bert-base-uncased"),
-        # ("longformer", "allenai/longformer-base-4096"),
+        ("bigbird", "google/bigbird-roberta-base"),
+        ("longformer", "allenai/longformer-base-4096"),
         # ("xlnet", "xlnet-base-cased"),
         # ("xlm", "xlm-mlm-17-1280"),
-        # ("roberta", "roberta-base"),
+        ("roberta", "roberta-base"),
         # ("distilbert", "distilbert-base-uncased"),
         # ("albert", "albert-base-v1"),
         # ("camembert", "camembert-base"),