use minicpm llama3 model

app.py

@@ -1,14 +1,17 @@
 import gradio as gr
 from huggingface_hub import InferenceClient
+import base64
 
-"""
-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
-"""
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
+# 更新为 MiniCPM-Llama3-V-2_5 模型
+client = InferenceClient("openbmb/MiniCPM-Llama3-V-2_5")
 
+def encode_image(image_path):
+    with open(image_path, "rb") as image_file:
+        return base64.b64encode(image_file.read()).decode('utf-8')
 
 def respond(
     message,
+    image,
     history: list[tuple[str, str]],
     system_message,
     max_tokens,
@@ -16,48 +19,47 @@ def respond(
     top_p,
 ):
     messages = [{"role": "system", "content": system_message}]
-
     for val in history:
         if val[0]:
             messages.append({"role": "user", "content": val[0]})
         if val[1]:
             messages.append({"role": "assistant", "content": val[1]})
+    
+    # 处理图片输入
+    if image:
+        base64_image = encode_image(image.name)
+        image_message = f"<image>{base64_image}</image>"
+        message = image_message + "\n" + message
 
     messages.append({"role": "user", "content": message})
-
+    
     response = ""
-
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
+    for message in client.text_generation(
+        prompt=f"{messages}",
+        max_new_tokens=max_tokens,
         stream=True,
         temperature=temperature,
         top_p=top_p,
     ):
-        token = message.choices[0].delta.content
-
+        token = message.token.text
         response += token
         yield response
 
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
-demo = gr.ChatInterface(
+demo = gr.Interface(
     respond,
-    additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
+    inputs=[
+        gr.Textbox(label="Message"),
+        gr.Image(type="filepath", label="Upload Image"),
+        gr.State([]),  # for history
+        gr.Textbox(value="You are a friendly AI assistant capable of understanding images and text.", 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)",
-        ),
+        gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p (nucleus sampling)"),
     ],
+    outputs=gr.Textbox(label="Response"),
+    title="MiniCPM-Llama3-V-2_5 Image and Text Chat",
+    description="Upload an image and ask questions about it, or just chat without an image."
 )
 
-
 if __name__ == "__main__":
     demo.launch()