Spaces:
				
			
			
	
			
			
		Running
		
			on 
			
			Zero
	
	
	
			
			
	
	
	
	
		
		
		Running
		
			on 
			
			Zero
	Update app.py
Browse files
    	
        app.py
    CHANGED
    
    | @@ -6,12 +6,11 @@ import os | |
| 6 | 
             
            import tempfile
         | 
| 7 | 
             
            from PIL import Image
         | 
| 8 |  | 
| 9 | 
            -
            # ---  | 
| 10 | 
            -
            # Di chuyển việc tải model ra ngoài để tránh tải lại mỗi lần gọi hàm
         | 
| 11 | 
             
            print("Loading model and tokenizer...")
         | 
| 12 | 
             
            model_name = "deepseek-ai/DeepSeek-OCR"
         | 
| 13 | 
             
            tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
         | 
| 14 | 
            -
            #  | 
| 15 | 
             
            model = AutoModel.from_pretrained(
         | 
| 16 | 
             
                model_name,
         | 
| 17 | 
             
                _attn_implementation="flash_attention_2",
         | 
| @@ -19,51 +18,50 @@ model = AutoModel.from_pretrained( | |
| 19 | 
             
                use_safetensors=True,
         | 
| 20 | 
             
            )
         | 
| 21 | 
             
            model = model.eval()
         | 
| 22 | 
            -
            print("Model loaded successfully.")
         | 
| 23 |  | 
| 24 |  | 
| 25 | 
            -
            # ---  | 
| 26 | 
             
            @spaces.GPU
         | 
| 27 | 
             
            def process_ocr_task(image, model_size, task_type, ref_text):
         | 
| 28 | 
             
                """
         | 
| 29 | 
            -
                 | 
| 30 | 
             
                Args:
         | 
| 31 | 
            -
                    image:  | 
| 32 | 
            -
                    model_size:  | 
| 33 | 
            -
                    task_type:  | 
| 34 | 
            -
                    ref_text:  | 
| 35 | 
             
                """
         | 
| 36 | 
             
                if image is None:
         | 
| 37 | 
             
                    return "Please upload an image first.", None
         | 
| 38 |  | 
| 39 | 
            -
                #  | 
| 40 | 
            -
                print("Moving model to GPU...")
         | 
| 41 | 
             
                model_gpu = model.cuda().to(torch.bfloat16)
         | 
| 42 | 
            -
                print("Model on GPU.")
         | 
| 43 |  | 
| 44 | 
            -
                #  | 
| 45 | 
             
                with tempfile.TemporaryDirectory() as output_path:
         | 
| 46 | 
            -
                    # ---  | 
| 47 | 
            -
                    if task_type == "Free OCR":
         | 
| 48 | 
             
                        prompt = "<image>\nFree OCR."
         | 
| 49 | 
            -
                    elif task_type == "Convert to Markdown":
         | 
| 50 | 
             
                        prompt = "<image>\n<|grounding|>Convert the document to markdown."
         | 
| 51 | 
            -
                    elif task_type == "Parse Figure":
         | 
| 52 | 
             
                        prompt = "<image>\nParse the figure."
         | 
| 53 | 
            -
                    elif task_type == "Locate Object by Reference":
         | 
| 54 | 
             
                        if not ref_text or ref_text.strip() == "":
         | 
| 55 | 
            -
                            raise gr.Error("For 'Locate' task,  | 
| 56 | 
            -
                        #  | 
| 57 | 
             
                        prompt = f"<image>\nLocate <|ref|>{ref_text.strip()}<|/ref|> in the image."
         | 
| 58 | 
             
                    else:
         | 
| 59 | 
            -
                         | 
| 60 | 
            -
                        prompt = "<image>\nFree OCR."
         | 
| 61 |  | 
| 62 | 
            -
                    #  | 
| 63 | 
             
                    temp_image_path = os.path.join(output_path, "temp_image.png")
         | 
| 64 | 
             
                    image.save(temp_image_path)
         | 
| 65 |  | 
| 66 | 
            -
                    #  | 
| 67 | 
             
                    size_configs = {
         | 
| 68 | 
             
                        "Tiny": {"base_size": 512, "image_size": 512, "crop_mode": False},
         | 
| 69 | 
             
                        "Small": {"base_size": 640, "image_size": 640, "crop_mode": False},
         | 
| @@ -73,8 +71,8 @@ def process_ocr_task(image, model_size, task_type, ref_text): | |
| 73 | 
             
                    }
         | 
| 74 | 
             
                    config = size_configs.get(model_size, size_configs["Gundam (Recommended)"])
         | 
| 75 |  | 
| 76 | 
            -
                    print(f"Running inference with prompt: {prompt}")
         | 
| 77 | 
            -
                    # ---  | 
| 78 | 
             
                    text_result = model_gpu.infer(
         | 
| 79 | 
             
                        tokenizer,
         | 
| 80 | 
             
                        prompt=prompt,
         | 
| @@ -83,120 +81,119 @@ def process_ocr_task(image, model_size, task_type, ref_text): | |
| 83 | 
             
                        base_size=config["base_size"],
         | 
| 84 | 
             
                        image_size=config["image_size"],
         | 
| 85 | 
             
                        crop_mode=config["crop_mode"],
         | 
| 86 | 
            -
                        save_results=True,  #  | 
| 87 | 
             
                        test_compress=True,
         | 
| 88 | 
             
                        eval_mode=True,
         | 
| 89 | 
             
                    )
         | 
| 90 |  | 
| 91 | 
            -
                    print(f"====\ | 
| 92 |  | 
| 93 | 
            -
                    # ---  | 
| 94 | 
             
                    image_result_path = None
         | 
| 95 | 
            -
                    #  | 
| 96 | 
            -
                    if task_type in ["Locate Object by Reference", "Convert to Markdown", "Parse Figure"]:
         | 
| 97 | 
            -
                        #  | 
| 98 | 
             
                        for filename in os.listdir(output_path):
         | 
| 99 | 
             
                            if "grounding" in filename or "result" in filename:
         | 
| 100 | 
             
                                image_result_path = os.path.join(output_path, filename)
         | 
| 101 | 
             
                                break
         | 
| 102 |  | 
| 103 | 
            -
                    #  | 
| 104 | 
             
                    result_image_pil = Image.open(image_result_path) if image_result_path else None
         | 
| 105 |  | 
| 106 | 
             
                    return text_result, result_image_pil
         | 
| 107 |  | 
| 108 |  | 
| 109 | 
            -
            # ---  | 
| 110 | 
            -
            with gr.Blocks(title="DeepSeek-OCR", theme=gr.themes.Soft()) as demo:
         | 
| 111 | 
             
                gr.Markdown(
         | 
| 112 | 
             
                    """
         | 
| 113 | 
            -
                    # Demo  | 
| 114 | 
            -
                     | 
| 115 |  | 
| 116 | 
            -
                     | 
| 117 | 
            -
                    1.  | 
| 118 | 
            -
                    2.  | 
| 119 | 
            -
                    3.  | 
| 120 | 
            -
                        -  | 
| 121 | 
            -
                        -  | 
| 122 | 
            -
                        -  | 
| 123 | 
            -
                        -  | 
| 124 | 
             
                    """
         | 
| 125 | 
             
                )
         | 
| 126 |  | 
| 127 | 
             
                with gr.Row():
         | 
| 128 | 
             
                    with gr.Column(scale=1):
         | 
| 129 | 
            -
                        image_input = gr.Image(type="pil", label=" | 
| 130 |  | 
| 131 | 
             
                        model_size = gr.Dropdown(
         | 
| 132 | 
             
                            choices=["Tiny", "Small", "Base", "Large", "Gundam (Recommended)"],
         | 
| 133 | 
             
                            value="Gundam (Recommended)",
         | 
| 134 | 
            -
                            label="Model Size",
         | 
| 135 | 
             
                        )
         | 
| 136 |  | 
| 137 | 
             
                        task_type = gr.Dropdown(
         | 
| 138 | 
            -
                            choices=["Free OCR", "Convert to Markdown", "Parse Figure", "Locate Object by Reference"],
         | 
| 139 | 
            -
                            value="Convert to Markdown",
         | 
| 140 | 
            -
                            label="Task Type",
         | 
| 141 | 
             
                        )
         | 
| 142 |  | 
| 143 | 
            -
                        # Ô nhập văn bản tham chiếu, ban đầu bị ẩn
         | 
| 144 | 
             
                        ref_text_input = gr.Textbox(
         | 
| 145 | 
            -
                            label="Reference Text ( | 
| 146 | 
            -
                            placeholder=" | 
| 147 | 
            -
                            visible=False, #  | 
| 148 | 
             
                        )
         | 
| 149 |  | 
| 150 | 
            -
                        submit_btn = gr.Button(" | 
| 151 |  | 
| 152 | 
             
                    with gr.Column(scale=2):
         | 
| 153 | 
            -
                        output_text = gr.Textbox(label=" | 
| 154 | 
            -
                        output_image = gr.Image(label=" | 
| 155 |  | 
| 156 | 
            -
                # ---  | 
| 157 | 
             
                def toggle_ref_text_visibility(task):
         | 
| 158 | 
            -
                    #  | 
| 159 | 
            -
                    if task == "Locate Object by Reference":
         | 
| 160 | 
             
                        return gr.Textbox(visible=True)
         | 
| 161 | 
             
                    else:
         | 
| 162 | 
             
                        return gr.Textbox(visible=False)
         | 
| 163 |  | 
| 164 | 
            -
                #  | 
| 165 | 
             
                task_type.change(
         | 
| 166 | 
             
                    fn=toggle_ref_text_visibility,
         | 
| 167 | 
             
                    inputs=task_type,
         | 
| 168 | 
             
                    outputs=ref_text_input,
         | 
| 169 | 
             
                )
         | 
| 170 |  | 
| 171 | 
            -
                #  | 
| 172 | 
             
                submit_btn.click(
         | 
| 173 | 
             
                    fn=process_ocr_task,
         | 
| 174 | 
             
                    inputs=[image_input, model_size, task_type, ref_text_input],
         | 
| 175 | 
             
                    outputs=[output_text, output_image],
         | 
| 176 | 
             
                )
         | 
| 177 |  | 
| 178 | 
            -
                # ---  | 
| 179 | 
             
                gr.Examples(
         | 
| 180 | 
             
                    examples=[
         | 
| 181 | 
            -
                        ["./examples/doc_markdown.png", "Gundam (Recommended)", "Convert to Markdown", ""],
         | 
| 182 | 
            -
                        ["./examples/chart.png", "Gundam (Recommended)", "Parse Figure", ""],
         | 
| 183 | 
            -
                        ["./examples/teacher.png", "Base", "Locate Object by Reference", "the teacher"],
         | 
| 184 | 
            -
                        ["./examples/math_locate.png", "Small", "Locate Object by Reference", "11-2="],
         | 
| 185 | 
            -
                        ["./examples/receipt.jpg", "Base", "Free OCR", ""],
         | 
| 186 | 
             
                    ],
         | 
| 187 | 
             
                    inputs=[image_input, model_size, task_type, ref_text_input],
         | 
| 188 | 
             
                    outputs=[output_text, output_image],
         | 
| 189 | 
             
                    fn=process_ocr_task,
         | 
| 190 | 
            -
                    cache_examples=False, #  | 
| 191 | 
             
                )
         | 
| 192 |  | 
| 193 | 
            -
            # ---  | 
| 194 | 
             
            if __name__ == "__main__":
         | 
| 195 | 
            -
                #  | 
| 196 | 
             
                if not os.path.exists("examples"):
         | 
| 197 | 
             
                    os.makedirs("examples")
         | 
| 198 | 
            -
                #  | 
| 199 | 
            -
                #  | 
| 200 |  | 
| 201 | 
             
                demo.queue(max_size=20)
         | 
| 202 | 
            -
                demo.launch(share=True) # share=True  | 
|  | |
| 6 | 
             
            import tempfile
         | 
| 7 | 
             
            from PIL import Image
         | 
| 8 |  | 
| 9 | 
            +
            # --- 1. Load Model and Tokenizer (Done only once at startup) ---
         | 
|  | |
| 10 | 
             
            print("Loading model and tokenizer...")
         | 
| 11 | 
             
            model_name = "deepseek-ai/DeepSeek-OCR"
         | 
| 12 | 
             
            tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
         | 
| 13 | 
            +
            # Load the model to CPU first; it will be moved to GPU during processing
         | 
| 14 | 
             
            model = AutoModel.from_pretrained(
         | 
| 15 | 
             
                model_name,
         | 
| 16 | 
             
                _attn_implementation="flash_attention_2",
         | 
|  | |
| 18 | 
             
                use_safetensors=True,
         | 
| 19 | 
             
            )
         | 
| 20 | 
             
            model = model.eval()
         | 
| 21 | 
            +
            print("✅ Model loaded successfully.")
         | 
| 22 |  | 
| 23 |  | 
| 24 | 
            +
            # --- 2. Main Processing Function ---
         | 
| 25 | 
             
            @spaces.GPU
         | 
| 26 | 
             
            def process_ocr_task(image, model_size, task_type, ref_text):
         | 
| 27 | 
             
                """
         | 
| 28 | 
            +
                Processes an image with DeepSeek-OCR for all supported tasks.
         | 
| 29 | 
             
                Args:
         | 
| 30 | 
            +
                    image (PIL.Image): The input image.
         | 
| 31 | 
            +
                    model_size (str): The model size configuration.
         | 
| 32 | 
            +
                    task_type (str): The type of OCR task to perform.
         | 
| 33 | 
            +
                    ref_text (str): The reference text for the 'Locate' task.
         | 
| 34 | 
             
                """
         | 
| 35 | 
             
                if image is None:
         | 
| 36 | 
             
                    return "Please upload an image first.", None
         | 
| 37 |  | 
| 38 | 
            +
                # Move the model to GPU and use bfloat16 for better performance
         | 
| 39 | 
            +
                print("🚀 Moving model to GPU...")
         | 
| 40 | 
             
                model_gpu = model.cuda().to(torch.bfloat16)
         | 
| 41 | 
            +
                print("✅ Model is on GPU.")
         | 
| 42 |  | 
| 43 | 
            +
                # Create a temporary directory to store files
         | 
| 44 | 
             
                with tempfile.TemporaryDirectory() as output_path:
         | 
| 45 | 
            +
                    # --- Build the prompt based on the selected task type ---
         | 
| 46 | 
            +
                    if task_type == "📝 Free OCR":
         | 
| 47 | 
             
                        prompt = "<image>\nFree OCR."
         | 
| 48 | 
            +
                    elif task_type == "📄 Convert to Markdown":
         | 
| 49 | 
             
                        prompt = "<image>\n<|grounding|>Convert the document to markdown."
         | 
| 50 | 
            +
                    elif task_type == "📈 Parse Figure":
         | 
| 51 | 
             
                        prompt = "<image>\nParse the figure."
         | 
| 52 | 
            +
                    elif task_type == "🔍 Locate Object by Reference":
         | 
| 53 | 
             
                        if not ref_text or ref_text.strip() == "":
         | 
| 54 | 
            +
                            raise gr.Error("For the 'Locate' task, you must provide the reference text to find!")
         | 
| 55 | 
            +
                        # Use an f-string to embed the user's reference text into the prompt
         | 
| 56 | 
             
                        prompt = f"<image>\nLocate <|ref|>{ref_text.strip()}<|/ref|> in the image."
         | 
| 57 | 
             
                    else:
         | 
| 58 | 
            +
                        prompt = "<image>\nFree OCR." # Default fallback
         | 
|  | |
| 59 |  | 
| 60 | 
            +
                    # Save the uploaded image to the temporary path
         | 
| 61 | 
             
                    temp_image_path = os.path.join(output_path, "temp_image.png")
         | 
| 62 | 
             
                    image.save(temp_image_path)
         | 
| 63 |  | 
| 64 | 
            +
                    # Configure model size parameters
         | 
| 65 | 
             
                    size_configs = {
         | 
| 66 | 
             
                        "Tiny": {"base_size": 512, "image_size": 512, "crop_mode": False},
         | 
| 67 | 
             
                        "Small": {"base_size": 640, "image_size": 640, "crop_mode": False},
         | 
|  | |
| 71 | 
             
                    }
         | 
| 72 | 
             
                    config = size_configs.get(model_size, size_configs["Gundam (Recommended)"])
         | 
| 73 |  | 
| 74 | 
            +
                    print(f"🏃 Running inference with prompt: {prompt}")
         | 
| 75 | 
            +
                    # --- Run the model's inference method ---
         | 
| 76 | 
             
                    text_result = model_gpu.infer(
         | 
| 77 | 
             
                        tokenizer,
         | 
| 78 | 
             
                        prompt=prompt,
         | 
|  | |
| 81 | 
             
                        base_size=config["base_size"],
         | 
| 82 | 
             
                        image_size=config["image_size"],
         | 
| 83 | 
             
                        crop_mode=config["crop_mode"],
         | 
| 84 | 
            +
                        save_results=True,  # Important: Must be True to get the output image
         | 
| 85 | 
             
                        test_compress=True,
         | 
| 86 | 
             
                        eval_mode=True,
         | 
| 87 | 
             
                    )
         | 
| 88 |  | 
| 89 | 
            +
                    print(f"====\n📄 Text Result: {text_result}\n====")
         | 
| 90 |  | 
| 91 | 
            +
                    # --- Handle the output (both text and image) ---
         | 
| 92 | 
             
                    image_result_path = None
         | 
| 93 | 
            +
                    # Tasks that generate a visual output usually create a 'grounding' or 'result' image
         | 
| 94 | 
            +
                    if task_type in ["🔍 Locate Object by Reference", "📄 Convert to Markdown", "📈 Parse Figure"]:
         | 
| 95 | 
            +
                        # Find the result image in the output directory
         | 
| 96 | 
             
                        for filename in os.listdir(output_path):
         | 
| 97 | 
             
                            if "grounding" in filename or "result" in filename:
         | 
| 98 | 
             
                                image_result_path = os.path.join(output_path, filename)
         | 
| 99 | 
             
                                break
         | 
| 100 |  | 
| 101 | 
            +
                    # If an image was found, open it with PIL; otherwise, return None
         | 
| 102 | 
             
                    result_image_pil = Image.open(image_result_path) if image_result_path else None
         | 
| 103 |  | 
| 104 | 
             
                    return text_result, result_image_pil
         | 
| 105 |  | 
| 106 |  | 
| 107 | 
            +
            # --- 3. Build the Gradio Interface ---
         | 
| 108 | 
            +
            with gr.Blocks(title="🐳DeepSeek-OCR🐳", theme=gr.themes.Soft()) as demo:
         | 
| 109 | 
             
                gr.Markdown(
         | 
| 110 | 
             
                    """
         | 
| 111 | 
            +
                    # 🐳 Full Demo of DeepSeek-OCR 🐳
         | 
| 112 | 
            +
                    Upload an image to explore the document recognition and understanding capabilities of DeepSeek-OCR.
         | 
| 113 |  | 
| 114 | 
            +
                    **💡 How to use:**
         | 
| 115 | 
            +
                    1.  **Upload an image** using the upload box.
         | 
| 116 | 
            +
                    2.  Select a **Model Size**. `Gundam` is recommended for most documents for a good balance of speed and accuracy.
         | 
| 117 | 
            +
                    3.  Choose a **Task Type**:
         | 
| 118 | 
            +
                        - **📝 Free OCR**: Extracts raw text from the image. Best for simple text extraction.
         | 
| 119 | 
            +
                        - **📄 Convert to Markdown**: Converts the entire document into Markdown format, preserving structure like headers, lists, and tables.
         | 
| 120 | 
            +
                        - **📈 Parse Figure**: Analyzes and extracts structured data from charts, graphs, and geometric figures.
         | 
| 121 | 
            +
                        - **🔍 Locate Object by Reference**: Finds a specific object or piece of text in the image. You **must** type what you're looking for into the **"Reference Text"** box that appears.
         | 
| 122 | 
             
                    """
         | 
| 123 | 
             
                )
         | 
| 124 |  | 
| 125 | 
             
                with gr.Row():
         | 
| 126 | 
             
                    with gr.Column(scale=1):
         | 
| 127 | 
            +
                        image_input = gr.Image(type="pil", label="🖼️ Upload Image", sources=["upload", "clipboard"])
         | 
| 128 |  | 
| 129 | 
             
                        model_size = gr.Dropdown(
         | 
| 130 | 
             
                            choices=["Tiny", "Small", "Base", "Large", "Gundam (Recommended)"],
         | 
| 131 | 
             
                            value="Gundam (Recommended)",
         | 
| 132 | 
            +
                            label="⚙️ Model Size",
         | 
| 133 | 
             
                        )
         | 
| 134 |  | 
| 135 | 
             
                        task_type = gr.Dropdown(
         | 
| 136 | 
            +
                            choices=["📝 Free OCR", "📄 Convert to Markdown", "📈 Parse Figure", "🔍 Locate Object by Reference"],
         | 
| 137 | 
            +
                            value="📄 Convert to Markdown",
         | 
| 138 | 
            +
                            label="🚀 Task Type",
         | 
| 139 | 
             
                        )
         | 
| 140 |  | 
|  | |
| 141 | 
             
                        ref_text_input = gr.Textbox(
         | 
| 142 | 
            +
                            label="📝 Reference Text (for Locate task)",
         | 
| 143 | 
            +
                            placeholder="e.g., the teacher, 11-2=, a red car...",
         | 
| 144 | 
            +
                            visible=False, # Initially hidden
         | 
| 145 | 
             
                        )
         | 
| 146 |  | 
| 147 | 
            +
                        submit_btn = gr.Button("Process Image", variant="primary")
         | 
| 148 |  | 
| 149 | 
             
                    with gr.Column(scale=2):
         | 
| 150 | 
            +
                        output_text = gr.Textbox(label="📄 Text Result", lines=15, show_copy_button=True)
         | 
| 151 | 
            +
                        output_image = gr.Image(label="🖼️ Image Result (if any)", type="pil")
         | 
| 152 |  | 
| 153 | 
            +
                # --- UI Interaction Logic ---
         | 
| 154 | 
             
                def toggle_ref_text_visibility(task):
         | 
| 155 | 
            +
                    # If the user selects the 'Locate' task, make the reference textbox visible
         | 
| 156 | 
            +
                    if task == "🔍 Locate Object by Reference":
         | 
| 157 | 
             
                        return gr.Textbox(visible=True)
         | 
| 158 | 
             
                    else:
         | 
| 159 | 
             
                        return gr.Textbox(visible=False)
         | 
| 160 |  | 
| 161 | 
            +
                # When the 'task_type' dropdown changes, call the function to update the visibility
         | 
| 162 | 
             
                task_type.change(
         | 
| 163 | 
             
                    fn=toggle_ref_text_visibility,
         | 
| 164 | 
             
                    inputs=task_type,
         | 
| 165 | 
             
                    outputs=ref_text_input,
         | 
| 166 | 
             
                )
         | 
| 167 |  | 
| 168 | 
            +
                # Define what happens when the submit button is clicked
         | 
| 169 | 
             
                submit_btn.click(
         | 
| 170 | 
             
                    fn=process_ocr_task,
         | 
| 171 | 
             
                    inputs=[image_input, model_size, task_type, ref_text_input],
         | 
| 172 | 
             
                    outputs=[output_text, output_image],
         | 
| 173 | 
             
                )
         | 
| 174 |  | 
| 175 | 
            +
                # --- Example Images and Tasks ---
         | 
| 176 | 
             
                gr.Examples(
         | 
| 177 | 
             
                    examples=[
         | 
| 178 | 
            +
                        ["./examples/doc_markdown.png", "Gundam (Recommended)", "📄 Convert to Markdown", ""],
         | 
| 179 | 
            +
                        ["./examples/chart.png", "Gundam (Recommended)", "📈 Parse Figure", ""],
         | 
| 180 | 
            +
                        ["./examples/teacher.png", "Base", "🔍 Locate Object by Reference", "the teacher"],
         | 
| 181 | 
            +
                        ["./examples/math_locate.png", "Small", "🔍 Locate Object by Reference", "11-2="],
         | 
| 182 | 
            +
                        ["./examples/receipt.jpg", "Base", "📝 Free OCR", ""],
         | 
| 183 | 
             
                    ],
         | 
| 184 | 
             
                    inputs=[image_input, model_size, task_type, ref_text_input],
         | 
| 185 | 
             
                    outputs=[output_text, output_image],
         | 
| 186 | 
             
                    fn=process_ocr_task,
         | 
| 187 | 
            +
                    cache_examples=False, # Disable caching to ensure examples run every time
         | 
| 188 | 
             
                )
         | 
| 189 |  | 
| 190 | 
            +
            # --- 4. Launch the App ---
         | 
| 191 | 
             
            if __name__ == "__main__":
         | 
| 192 | 
            +
                # Create an 'examples' directory if it doesn't exist
         | 
| 193 | 
             
                if not os.path.exists("examples"):
         | 
| 194 | 
             
                    os.makedirs("examples")
         | 
| 195 | 
            +
                # Please manually download the example images into the "examples" folder.
         | 
| 196 | 
            +
                # e.g., doc_markdown.png, chart.png, teacher.png, math_locate.png, receipt.jpg
         | 
| 197 |  | 
| 198 | 
             
                demo.queue(max_size=20)
         | 
| 199 | 
            +
                demo.launch(share=True) # Set share=True to create a public link
         | 
 
			
