[CodeGen] Add CodeGen model for code generation (#2641)

* feat: add CodeGen model

* update copyright

Author: gongenlei

paddlenlp/transformers/__init__.py

@@ -114,6 +114,8 @@
 from .optimization import *
 from .auto.modeling import *
 from .auto.tokenizer import *
+from .codegen.modeling import *
+from .codegen.tokenizer import *
 
 # For faster tokenizer
 from ..utils.import_utils import is_faster_tokenizer_available

paddlenlp/transformers/auto/modeling.py

@@ -78,6 +78,7 @@
     ("Bert", "bert"),
     ("Bart", "bart"),
     ("GAUAlpha", "gau_alpha"),
+    ("CodeGen", "codegen"),
 ])
 
 MAPPING_TASKS = OrderedDict([

paddlenlp/transformers/auto/tokenizer.py

@@ -77,6 +77,7 @@
     ("BertTokenizer", "bert"),
     ("BartTokenizer", "bart"),
     ("GAUAlphaTokenizer", "gau_alpha"),
+    ("CodeGenTokenizer", "codegen"),
 ])
 
 FASTER_TOKENIZER_MAPPING_NAMES = OrderedDict([("BertFasterTokenizer", "bert"),