test

app.py

@@ -1,70 +1,103 @@
 import gradio as gr
-from huggingface_hub import InferenceClient
+import torch
+from PIL import Image
+import requests
+from io import BytesIO
+from transformers import Qwen3VLForConditionalGeneration, AutoProcessor
 
+# --- Configuration ---
+MODEL_PATH = "zhangbaoxin/qwen3-vl-2b-package_unsloth_finetune"
+CPU_DEVICE = "cpu"
 
-def respond(
-    message,
-    history: list[dict[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-    hf_token: gr.OAuthToken,
-):
+# --- Model and Processor Loading ---
+print("Loading model and processor... This will take a few minutes on a CPU.")
+processor = AutoProcessor.from_pretrained(MODEL_PATH, trust_remote_code=True)
+model = Qwen3VLForConditionalGeneration.from_pretrained(
+    MODEL_PATH,
+    trust_remote_code=True,
+    dtype="auto",  # Use 'auto' for dtype for better compatibility
+    device_map="auto" # This is the key for CPU (and GPU) compatibility
+)
+print("Model and processor loaded successfully.")
+
+# --- Inference Function ---
+def process_and_generate(image_input, text_prompt):
     """
-    For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
+    Processes the image and text prompt, and generates a response from the model.
     """
-    client = InferenceClient(token=hf_token.token, model="openai/gpt-oss-20b")
+    if image_input is None or not text_prompt.strip():
+        return "Please provide both an image and a text prompt."
 
-    messages = [{"role": "system", "content": system_message}]
+    # Convert Gradio's numpy array to a PIL Image
+    pil_image = Image.fromarray(image_input)
 
-    messages.extend(history)
+    # Prepare the messages payload for the model
+    messages = [
+        {
+            "role": "user",
+            "content": [
+                {"type": "image", "image": pil_image},
+                {"type": "text", "text": text_prompt},
+            ],
+        }
+    ]
 
-    messages.append({"role": "user", "content": message})
+    print("Processing inputs and generating response... This will be slow.")
+    try:
+        # Preparation for inference
+        inputs = processor.apply_chat_template(
+            messages,
+            tokenize=True,
+            add_generation_prompt=True,
+            return_dict=True,
+            return_tensors="pt"
+        )
+        inputs = inputs.to(model.device)
 
-    response = ""
+        # Inference: Generation of the output
+        generated_ids = model.generate(**inputs, max_new_tokens=1024)
 
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        choices = message.choices
-        token = ""
-        if len(choices) and choices[0].delta.content:
-            token = choices[0].delta.content
+        # To get only the new tokens, we trim the input IDs from the generated IDs
+        generated_ids_trimmed = [
+            out_ids[len(in_ids):] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
+        ]
+        
+        # Decode the trimmed IDs to text
+        output_text = processor.batch_decode(
+            generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
+        )
+        
+        # batch_decode returns a list, we return the first element
+        return output_text[0]
+        
+    except Exception as e:
+        return f"An error occurred during generation: {str(e)}"
 
-        response += token
-        yield response
+# --- Gradio Interface ---
+with gr.Blocks() as demo:
+    gr.Markdown(
+        """
+        # Qwen3-VL-2B-Instruct CPU Demo
+        This Space runs the `Qwen/Qwen3-VL-2B-Instruct` model using the standard `transformers` library.
+        **Warning:** Running this on a free CPU Space is **very slow**. Duplicate this space for solo experience.
+        """
+    )
 
+    with gr.Row():
+        with gr.Column():
+            image_input = gr.Image(type="numpy", label="Upload Image")
+            text_prompt = gr.Textbox(label="Prompt", placeholder="e.g., Describe this image in detail.")
+            submit_button = gr.Button("Generate Response")
+        with gr.Column():
+            output_text = gr.Textbox(label="Model Output", lines=10, interactive=False)
 
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
-chatbot = gr.ChatInterface(
-    respond,
-    type="messages",
-    additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(
-            minimum=0.1,
-            maximum=1.0,
-            value=0.95,
-            step=0.05,
-            label="Top-p (nucleus sampling)",
-        ),
-    ],
-)
+    submit_button.click(
+        fn=process_and_generate,
+        inputs=[image_input, text_prompt],
+        outputs=output_text
+    )
 
-with gr.Blocks() as demo:
-    with gr.Sidebar():
-        gr.LoginButton()
-    chatbot.render()
 
 
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch()