test new path

global_local/models/video_instruction_ft_model.py

@@ -30,12 +30,6 @@ class VideoInstructionFTLLAMA(Blip2Base):
 
     @classmethod
     def init_video_Qformer(cls, num_query_token, vision_width,num_hidden_layers =2):
-
-        model = AutoModel.from_pretrained("bert-base-uncased")
-        model.push_to_hub("bert-base-uncased")
-        #model.push_to_hub("huggingface/bert-base-uncased")
-        sys.exit()
-
         encoder_config = BertConfig.from_pretrained("bert-base-uncased")
         encoder_config.num_hidden_layers = num_hidden_layers
         encoder_config.encoder_width = vision_width
@@ -83,6 +77,12 @@ class VideoInstructionFTLLAMA(Blip2Base):
         equip_audio_branch = True
     ):
         super().__init__()
+        tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
+        model = AutoModel.from_pretrained("bert-base-uncased")
+        model.push_to_hub("bert-base-uncased")
+        tokenizer.push_to_hub("bert-base-uncased")
+        #model.push_to_hub("huggingface/bert-base-uncased")
+        sys.exit()
 
         self.tokenizer = self.init_tokenizer()
         self.low_resource = low_resource