Update app.py

app.py

@@ -6,14 +6,21 @@ torch.jit.script = lambda f: f  # Avoid script error in lambda
 from t2v_metrics import VQAScore, list_all_vqascore_models
 
 # Global model variable, but do not initialize or move to CUDA here
-model_pipe = VQAScore(model="clip-flant5-xl", device="cuda")  # our recommended scoring model
+cur_model_name = "clip-flant5-xl"
+model_pipe = update_model(cur_model_name)
+
+def update_model(model_name):
+    if model_nm
+    return VQAScore(model=model_name, device="cuda") 
 
 @spaces.GPU(duration = 20)
 def generate(model_name, image, text):
-    print(list_all_vqascore_models())  # Debug: List available models
+    if model_name != cur_model_name:
+        model_pipe = update_model(model_name)
+    
     print("Image:", image)  # Debug: Print image path
     print("Text:", text)  # Debug: Print text input
-    print("Generating!")
+    print("Using model:", model_name)
     # Wrap the model call in a try-except block to capture and debug CUDA errors
     try:
         result = model_pipe(images=[image], texts=[text]).cpu()[0][0].item()  # Perform the model inference
@@ -26,6 +33,7 @@ def generate(model_name, image, text):
 
 demo = gr.Interface(
     fn=generate,  # function to call
+    # ['clip-flant5-xxl', 'clip-flant5-xl', 'clip-flant5-xxl-no-system', 'clip-flant5-xxl-no-system-no-user', 'llava-v1.5-13b', 'llava-v1.5-7b', 'sharegpt4v-7b', 'sharegpt4v-13b', 'llava-v1.6-13b', 'instructblip-flant5-xxl', 'instructblip-flant5-xl']
     inputs=[gr.Dropdown(["clip-flant5-xl", "clip-flant5-xxl"], label="Model Name"), gr.Image(type="filepath"), gr.Textbox(label="Prompt")],  # define the types of inputs
     outputs="number",  # define the type of output
     title="VQAScore",  # title of the app