update app

app.py

@@ -138,6 +138,24 @@ model_y = Qwen3VLForConditionalGeneration.from_pretrained(
     torch_dtype=torch.float16
 ).to(device).eval()
 
+# Load Qwen3-VL-2B-Instruct
+MODEL_ID_L = "Qwen/Qwen3-VL-2B-Instruct"
+processor_l = AutoProcessor.from_pretrained(MODEL_ID_L, trust_remote_code=True)
+model_l = Qwen3VLForConditionalGeneration.from_pretrained(
+    MODEL_ID_L,
+    trust_remote_code=True,
+    torch_dtype=torch.float16
+).to(device).eval()
+
+# Load Qwen3-VL-2B-Thinking
+MODEL_ID_J = "Qwen/Qwen3-VL-2B-Thinking"
+processor_j = AutoProcessor.from_pretrained(MODEL_ID_J, trust_remote_code=True)
+model_j = Qwen3VLForConditionalGeneration.from_pretrained(
+    MODEL_ID_J,
+    trust_remote_code=True,
+    torch_dtype=torch.float16
+).to(device).eval()
+
 # Load Qwen3-VL-4B-Thinking
 MODEL_ID_T = "Qwen/Qwen3-VL-4B-Thinking"
 processor_t = AutoProcessor.from_pretrained(MODEL_ID_T, trust_remote_code=True)
@@ -189,6 +207,10 @@ def generate_image(model_name: str, text: str, image: Image.Image,
         processor, model = processor_y, model_y
     elif model_name == "Qwen3-VL-4B-Thinking":
         processor, model = processor_t, model_t
+    elif model_name == "Qwen3-VL-2B-Instruct":
+        processor, model = processor_l, model_l
+    elif model_name == "Qwen3-VL-2B-Thinking":
+        processor, model = processor_j, model_j      
     else:
         yield "Invalid model selected.", "Invalid model selected."
         return
@@ -231,6 +253,10 @@ def generate_video(model_name: str, text: str, video_path: str,
         processor, model = processor_y, model_y
     elif model_name == "Qwen3-VL-4B-Thinking":
         processor, model = processor_t, model_t
+    elif model_name == "Qwen3-VL-2B-Instruct":
+        processor, model = processor_l, model_l
+    elif model_name == "Qwen3-VL-2B-Thinking":
+        processor, model = processor_j, model_j   
     else:
         yield "Invalid model selected.", "Invalid model selected."
         return
@@ -275,8 +301,8 @@ image_examples = [
     ["Explain the content (ocr).", "images/O.jpg"],
     ["What is the core meaning of the poem?", "images/S.jpg"],
     ["Provide a detailed caption for the image.", "images/A.jpg"],
-    ["Explain the pie-chart in detail.", "images/2.jpg"],
-    ["Jsonify Data.", "images/1.jpg"],
+    #["Explain the pie-chart in detail.", "images/2.jpg"],
+    #["Jsonify Data.", "images/1.jpg"],
 ]
 
 video_examples = [
@@ -325,7 +351,7 @@ with gr.Blocks(css=css, theme=steel_blue_theme) as demo:
                 markdown_output = gr.Markdown()
                 
             model_choice = gr.Radio(
-                choices=["Qwen3-VL-4B-Instruct", "Qwen3-VL-8B-Instruct", "Qwen3-VL-4B-Thinking", "Qwen2.5-VL-3B-Instruct", "Qwen2.5-VL-7B-Instruct"],
+                choices=["Qwen3-VL-4B-Instruct", "Qwen3-VL-8B-Instruct",  "Qwen3-VL-2B-Instruct", "Qwen3-VL-2B-Thinking", "Qwen3-VL-4B-Thinking", "Qwen2.5-VL-3B-Instruct", "Qwen2.5-VL-7B-Instruct"],
                 label="Select Model",
                 value="Qwen3-VL-4B-Instruct"
             )