Update app.py

app.py

@@ -84,13 +84,15 @@ def chat_with_model(image_path=None, text_query=None, history=None):
         return "An error occurred while processing your request."
 
 def draw_boxes(image_path, response):
-    image = Image.open(image_path)
-    draw = ImageDraw.Draw(image)
-    boxes = re.findall(r'<box>\((\d+),(\d+)\),\((\d+),(\d+)\)</box>', response)
-    for box in boxes:
-        x1, y1, x2, y2 = map(int, box)
-        draw.rectangle([x1, y1, x2, y2], outline="red", width=3)
-    return image
+    with Image.open(image_path) as image:
+        draw = ImageDraw.Draw(image)
+        boxes = re.findall(r'<box>\((\d+),(\d+)\),\((\d+),(\d+)\)</box>', response)
+        for box in boxes:
+            x1, y1, x2, y2 = map(int, box)
+            draw.rectangle([x1, y1, x2, y2], outline="red", width=3)
+        with tempfile.NamedTemporaryFile(delete=False, suffix='.png', dir=os.path.abspath("uploaded_images")) as temp_file:
+            image.save(temp_file, format="PNG")
+            return temp_file.name
 
 def process_input(text=None, file=None, task_history=None):
     if task_history is None:
@@ -105,11 +107,8 @@ def process_input(text=None, file=None, task_history=None):
     task_history.append((text, response))
 
     if "<box>" in response and image_path:
-        image_with_boxes = draw_boxes(image_path, response)
-        buffered = BytesIO()
-        image_with_boxes.save(buffered, format="PNG")
-        img_str = buffered.getvalue()
-        return [("bot", response, "data:image/png;base64," + base64.b64encode(img_str).decode())], task_history
+        image_with_boxes_path = draw_boxes(image_path, response)
+        return [("bot", response), "image", image_with_boxes_path], task_history
     else:
         cleaned_response = clean_response(response)
         return [("bot", cleaned_response)], task_history
@@ -117,11 +116,11 @@ def process_input(text=None, file=None, task_history=None):
 with gr.Blocks() as demo:
     gr.Markdown("""
 # 🙋🏻‍♂️欢迎来到🌟Tonic 的🦆Qwen-VL-Chat🤩Bot！🚀
-# 🙋🏻‍♂️Welcome toTonic's Qwen-VL-Chat Bot! 
+# 🙋🏻‍♂️Welcome to Tonic's🦆Qwen-VL-Chat🤩Bot！🚀
 该WebUI基于Qwen-VL-Chat，实现聊天机器人功能。 但我必须解决它的很多问题，也许我也能获得一些荣誉。
 Qwen-VL-Chat 是一种多模式输入模型。 您可以使用此空间来测试当前模型 [qwen/Qwen-VL-Chat](https://huggingface.co/qwen/Qwen-VL-Chat) 您也可以使用 🧑🏻‍🚀qwen/Qwen-VL -通过克隆这个空间来聊天🚀。 🧬🔬🔍 只需点击这里：[重复空间](https://huggingface.co/spaces/Tonic1/VLChat?duplicate=true)
 加入我们：🌟TeamTonic🌟总是在制作很酷的演示！ 在 👻Discord 上加入我们活跃的构建者🛠️社区：[Discord](https://discord.gg/nXx5wbX9) 在 🤗Huggingface 上：[TeamTonic](https://huggingface.co/TeamTonic) 和 [MultiTransformer](https:/ /huggingface.co/MultiTransformer) 在 🌐Github 上：[Polytonic](https://github.com/tonic-ai) 并为 🌟 [PolyGPT](https://github.com/tonic-ai/polygpt-alpha) 做出贡献 ）
-This WebUI is based on Qwen-VL-Chat, implementing chatbot functionalities. Qwen-VL-Chat is a multimodal input model. You can use this Space to test out the current model [qwen/Qwen-VL-Chat](https://huggingface.co/qwen/Qwen-VL-Chat) You can also use  qwen/Qwen-VL-Chat🚀 by cloning this space.   Simply click here: [Duplicate Space](https://huggingface.co/spaces/Tonic1/VLChat?duplicate=true)
+**You can upload a picture with query or specifically ask this model to draw boxes and frames (just keep trying until you get your returns)** . Qwen-VL-Chat is a multimodal input model. You can use this Space to test out the current model [qwen/Qwen-VL-Chat](https://huggingface.co/qwen/Qwen-VL-Chat) You can also use  qwen/Qwen-VL-Chat🚀 by cloning this space.   Simply click here: [Duplicate Space](https://huggingface.co/spaces/Tonic1/VLChat?duplicate=true)
 Join us:  TeamTonic  is always making cool demos! Join our active builder's community on  Discord: [Discord](https://discord.gg/nXx5wbX9) On  Huggingface: [TeamTonic](https://huggingface.co/TeamTonic) & [MultiTransformer](https://huggingface.co/MultiTransformer) On  Github: [Polytonic](https://github.com/tonic-ai) & contribute to   [PolyGPT](https://github.com/tonic-ai/polygpt-alpha)
 """)
     with gr.Row():