[Entrypoints] Support Qwen-Image-Edit 2509 online serving & update doc (#368)

Signed-off-by: David Chen <530634352@qq.com>

examples/offline_inference/image_to_image/image_to_image.md

@@ -15,7 +15,7 @@ python image_edit.py \
   --cfg_scale 4.0 \
 ```
 
-### Multiple Image Editing (Qwen-Image-Edit-2509+)
+### Multiple Image Editing (Qwen-Image-Edit-2509)
 
 For multiple image inputs, use `Qwen/Qwen-Image-Edit-2509` or later version:
 

examples/online_serving/image_to_image/README.md

@@ -2,6 +2,8 @@
 
 This example demonstrates how to deploy Qwen-Image-Edit model for online image editing service using vLLM-Omni.
 
+For **multi-image** input editing, use **Qwen-Image-Edit-2509** (QwenImageEditPlusPipeline) and send multiple images in the user message content.
+
 ## Start Server
 
 ### Basic Start
@@ -10,6 +12,12 @@ This example demonstrates how to deploy Qwen-Image-Edit model for online image e
 vllm serve Qwen/Qwen-Image-Edit --omni --port 8092
 ```
 
+### Multi-Image Edit (Qwen-Image-Edit-2509)
+
+```bash
+vllm serve Qwen/Qwen-Image-Edit-2509 --omni --port 8092
+```
+
 ### Start with Parameters
 
 
@@ -19,6 +27,12 @@ Or use the startup script:
 bash run_server.sh
 ```
 
+To serve Qwen-Image-Edit-2509 with the script:
+
+```bash
+MODEL=Qwen/Qwen-Image-Edit-2509 bash run_server.sh
+```
+
 ## API Calls
 
 ### Method 1: Using curl (Image Editing)
@@ -53,6 +67,9 @@ curl -s http://localhost:8092/v1/chat/completions \
 
 ```bash
 python openai_chat_client.py --input input.png --prompt "Convert to oil painting style" --output output.png
+
+# Multi-image editing (Qwen-Image-Edit-2509 server required)
+python openai_chat_client.py --input input1.png input2.png --prompt "Combine these images into a single scene" --output output.png
 ```
 
 ### Method 3: Using Gradio Demo
@@ -121,6 +138,25 @@ Use `extra_body` to pass generation parameters:
 }
 ```
 
+### Multi-Image Editing (Qwen-Image-Edit-2509)
+
+Provide multiple images in `content` (order matters):
+
+```json
+{
+  "messages": [
+    {
+      "role": "user",
+      "content": [
+        {"type": "text", "text": "Combine these images into a single scene"},
+        {"type": "image_url", "image_url": {"url": "data:image/png;base64,..."} },
+        {"type": "image_url", "image_url": {"url": "data:image/png;base64,..."} }
+      ]
+    }
+  ]
+}
+```
+
 ## Generation Parameters (extra_body)
 
 | Parameter                | Type  | Default | Description                           |

examples/online_serving/image_to_image/gradio_demo.py

@@ -15,8 +15,15 @@ import requests
 from PIL import Image
 
 
+def _pil_to_b64_png(img: Image.Image) -> str:
+    buffer = BytesIO()
+    img.save(buffer, format="PNG")
+    return base64.b64encode(buffer.getvalue()).decode("utf-8")
+
+
 def edit_image(
     input_image: Image.Image,
+    extra_images: list[str] | None,
     prompt: str,
     steps: int,
     guidance_scale: float,
@@ -28,19 +35,23 @@ def edit_image(
     if input_image is None:
         raise gr.Error("Please upload an image first")
 
-    # Convert image to base64
-    buffer = BytesIO()
-    input_image.save(buffer, format="PNG")
-    image_b64 = base64.b64encode(buffer.getvalue()).decode("utf-8")
+    images: list[Image.Image] = [input_image]
+    if extra_images:
+        for p in extra_images:
+            try:
+                images.append(Image.open(p).convert("RGB"))
+            except Exception as e:
+                raise gr.Error(f"Failed to open image: {p}. Error: {e}") from e
 
     # Build user message with text and image
+    content: list[dict[str, object]] = [{"type": "text", "text": prompt}]
+    for img in images:
+        content.append({"type": "image_url", "image_url": {"url": f"data:image/png;base64,{_pil_to_b64_png(img)}"}})
+
     messages = [
         {
             "role": "user",
-            "content": [
-                {"type": "text", "text": prompt},
-                {"type": "image_url", "image_url": {"url": f"data:image/png;base64,{image_b64}"}},
-            ],
+            "content": content,
         }
     ]
 
@@ -87,7 +98,10 @@ def create_demo(server_url: str):
 
     with gr.Blocks(title="Qwen-Image-Edit Demo") as demo:
         gr.Markdown("# Qwen-Image-Edit Online Editing")
-        gr.Markdown("Upload an image and describe the editing effect you want")
+        gr.Markdown(
+            "Upload an image and describe the editing effect you want. "
+            "For multi-image editing, upload extra images (requires Qwen-Image-Edit-2509 server)."
+        )
 
         with gr.Row():
             with gr.Column(scale=1):
@@ -95,6 +109,11 @@ def create_demo(server_url: str):
                     label="Input Image",
                     type="pil",
                 )
+                extra_images = gr.File(
+                    label="Additional Images (Optional)",
+                    file_count="multiple",
+                    type="filepath",
+                )
                 prompt = gr.Textbox(
                     label="Edit Instruction",
                     placeholder="Describe the editing effect you want...",
@@ -152,13 +171,13 @@ def create_demo(server_url: str):
             inputs=[prompt],
         )
 
-        def process_edit(img, p, st, g, se, n):
+        def process_edit(img, imgs, p, st, g, se, n):
             actual_seed = se if se >= 0 else None
-            return edit_image(img, p, st, g, actual_seed, n, server_url)
+            return edit_image(img, imgs, p, st, g, actual_seed, n, server_url)
 
         edit_btn.click(
             fn=process_edit,
-            inputs=[input_image, prompt, steps, guidance_scale, seed, negative_prompt],
+            inputs=[input_image, extra_images, prompt, steps, guidance_scale, seed, negative_prompt],
             outputs=[output_image],
         )
 

examples/online_serving/image_to_image/openai_chat_client.py

@@ -5,6 +5,7 @@ Qwen-Image-Edit OpenAI-compatible chat client for image editing.
 Usage:
     python openai_chat_client.py --input qwen_image_output.png --prompt "Convert to watercolor style" --output output.png
     python openai_chat_client.py --input input.png --prompt "Convert to oil painting" --seed 42
+    python openai_chat_client.py --input input1.png input2.png --prompt "Combine these images into a single scene"
 """
 
 import argparse
@@ -16,8 +17,19 @@ import requests
 from PIL import Image
 
 
+def _encode_image_as_data_url(input_path: Path) -> str:
+    image_bytes = input_path.read_bytes()
+    try:
+        img = Image.open(BytesIO(image_bytes))
+        mime_type = f"image/{img.format.lower()}" if img.format else "image/png"
+    except Exception:
+        mime_type = "image/png"
+    image_b64 = base64.b64encode(image_bytes).decode("utf-8")
+    return f"data:{mime_type};base64,{image_b64}"
+
+
 def edit_image(
-    input_image: str | Path,
+    input_image: str | Path | list[str | Path],
     prompt: str,
     server_url: str = "http://localhost:8092",
     height: int | None = None,
@@ -30,7 +42,7 @@ def edit_image(
     """Edit an image using the chat completions API.
 
     Args:
-        input_image: Path to input image
+        input_image: Path(s) to input image(s). For multi-image editing, pass multiple paths.
         prompt: Text description of the edit
         server_url: Server URL
         height: Output image height in pixels
@@ -43,28 +55,22 @@ def edit_image(
     Returns:
         Edited image bytes or None if failed
     """
-    # Read and encode input image
-    input_path = Path(input_image)
-    if not input_path.exists():
-        print(f"Error: Input image not found: {input_path}")
-        return None
-
-    with open(input_path, "rb") as f:
-        image_bytes = f.read()
-    image_b64 = base64.b64encode(image_bytes).decode("utf-8")
-
-    # Detect image type
-    img = Image.open(BytesIO(image_bytes))
-    mime_type = f"image/{img.format.lower()}" if img.format else "image/png"
+    input_images = input_image if isinstance(input_image, list) else [input_image]
+    input_paths = [Path(p) for p in input_images]
+    for p in input_paths:
+        if not p.exists():
+            print(f"Error: Input image not found: {p}")
+            return None
 
     # Build user message with text and image
+    content: list[dict[str, object]] = [{"type": "text", "text": prompt}]
+    for p in input_paths:
+        content.append({"type": "image_url", "image_url": {"url": _encode_image_as_data_url(p)}})
+
     messages = [
         {
             "role": "user",
-            "content": [
-                {"type": "text", "text": prompt},
-                {"type": "image_url", "image_url": {"url": f"data:{mime_type};base64,{image_b64}"}},
-            ],
+            "content": content,
         }
     ]
 
@@ -113,7 +119,7 @@ def edit_image(
 
 def main():
     parser = argparse.ArgumentParser(description="Qwen-Image-Edit chat client")
-    parser.add_argument("--input", "-i", required=True, help="Input image path")
+    parser.add_argument("--input", "-i", required=True, nargs="+", help="Input image path(s)")
     parser.add_argument("--prompt", "-p", required=True, help="Edit prompt")
     parser.add_argument("--output", "-o", default="output.png", help="Output file")
     parser.add_argument("--server", "-s", default="http://localhost:8092", help="Server URL")
@@ -126,7 +132,10 @@ def main():
 
     args = parser.parse_args()
 
-    print(f"Input: {args.input}")
+    if len(args.input) == 1:
+        print(f"Input: {args.input[0]}")
+    else:
+        print(f"Inputs ({len(args.input)}): {', '.join(args.input)}")
     print(f"Prompt: {args.prompt}")
 
     image_bytes = edit_image(