* Add StarChat2

* Add DPO

* Fix unit test

* Typos

* Typo
This commit is contained in:
lewtun
2024-03-12 17:22:21 +01:00
committed by GitHub
parent ff618a4d13
commit a9b8a50a27
14 changed files with 324 additions and 29 deletions
+2 -2
View File
@@ -70,8 +70,8 @@ class GetTokenizerTest(unittest.TestCase):
`default_chat_template` but no `chat_template` we do not set a `chat_template`,
and that we do not change `default_chat_template`
"""
model_args = ModelArguments(model_name_or_path="codellama/CodeLlama-7b-Instruct-hf")
base_tokenizer = AutoTokenizer.from_pretrained("codellama/CodeLlama-7b-Instruct-hf")
model_args = ModelArguments(model_name_or_path="m-a-p/OpenCodeInterpreter-SC2-7B")
base_tokenizer = AutoTokenizer.from_pretrained("m-a-p/OpenCodeInterpreter-SC2-7B")
processed_tokenizer = get_tokenizer(model_args, DataArguments())
assert getattr(processed_tokenizer, "chat_template") is None