Update app.py

app.py

@@ -1,6 +1,6 @@
 import gradio as gr
 import torch
-from transformers import AutoModel, AutoTokenizer
+from transformers import AutoModel, AutoTokenizer, AutoModelForCasualLM
 import deepspeed
 
 # Model name
@@ -10,12 +10,12 @@ model_name = "OpenGVLab/InternVideo2_5_Chat_8B"
 tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
 
 # Load model efficiently
-model = AutoModel.from_pretrained(
+model = AutoModelForCasualLM.from_pretrained(
     model_name, 
     trust_remote_code=True,
     torch_dtype=torch.float16,  # Use float16 for lower memory usage
     device_map="auto", # Automatically place model on available GPU
-    deepspeed={"stage": 3}
+    
 )
 
 # Define inference function