diff --git a/modelscope/preprocessors/ofa/summarization.py b/modelscope/preprocessors/ofa/summarization.py index 8568a543..d33e9d25 100644 --- a/modelscope/preprocessors/ofa/summarization.py +++ b/modelscope/preprocessors/ofa/summarization.py @@ -46,7 +46,6 @@ class OfaSummarizationPreprocessor(OfaBasePreprocessor): def _build_infer_sample(self, data: Dict[str, Any]) -> Dict[str, Any]: source = super().pre_caption( data[self.column_map['text']], max_words=self.max_src_length) - # source = source.strip()[:self.max_src_length] source = source.replace('[unk]', 'unk').replace('', 'unk') prompt = self.cfg.model.get( 'prompt', ' " {} " Summarize the article with a title: ') diff --git a/tests/trainers/test_ofa_trainer.py b/tests/trainers/test_ofa_trainer.py index 098416bb..f72400eb 100644 --- a/tests/trainers/test_ofa_trainer.py +++ b/tests/trainers/test_ofa_trainer.py @@ -71,13 +71,20 @@ class TestOfaTrainer(unittest.TestCase): @unittest.skipUnless(test_level() >= 0, 'skip test in current test level') def test_trainer_std(self): + # WORKSPACE = './workspace/ckpts/recognition' + # os.makedirs(WORKSPACE, exist_ok=True) + # + # pretrained_model = 'damo/ofa_ocr-recognition_scene_base_zh' + # cfg = read_config(pretrained_model) + # config_file = os.path.join(WORKSPACE, ModelFile.CONFIGURATION) + # cfg.dump(config_file) WORKSPACE = './workspace/ckpts/recognition' os.makedirs(WORKSPACE, exist_ok=True) + config_file = os.path.join(WORKSPACE, ModelFile.CONFIGURATION) + with open(config_file, 'w') as writer: + json.dump(self.finetune_cfg, writer) pretrained_model = 'damo/ofa_ocr-recognition_scene_base_zh' - cfg = read_config(pretrained_model) - config_file = os.path.join(WORKSPACE, ModelFile.CONFIGURATION) - cfg.dump(config_file) args = dict( model=pretrained_model,