Vision-CAIR
/

MiniGPT4-video-llama-hf

Feature Extraction

8-bit precision

Model card Files Files and versions Community

Vision-CAIR commited on Jul 24

Commit

2202d00

•

1 Parent(s): 87cd64a

Push model using huggingface_hub.

Files changed (1) hide show

mini_gpt4_llama_v2.py +2 -2

mini_gpt4_llama_v2.py CHANGED Viewed

@@ -88,8 +88,8 @@ class MiniGPT4_Video(Blip2Base, PreTrainedModel):
         cfg={},
     ):
         ## loop through the config minigpt4_video_config object and set the attributes
-        if isinstance(cfg, minigpt4_video_config):
-            cfg = cfg.to_dict()
         for key, value in cfg.items():
             try:

         cfg={},
     ):
         ## loop through the config minigpt4_video_config object and set the attributes
+        # if isinstance(cfg, minigpt4_video_config):
+        cfg = cfg.to_dict()
         for key, value in cfg.items():
             try: