init app

app.py

@@ -0,0 +1,113 @@
+# app.py
+from transformers import AutoTokenizer, AutoModelForCausalLM, TextIteratorStreamer
+from threading import Thread
+import gradio as gr
+import re
+import torch
+from openai import OpenAI
+
+#client = OpenAI(
+#    api_key="sk-420ab66020704eabbe37501ec39b7a2b",
+#    base_url="https://bailingchat.alipay.com",
+#)
+
+client = OpenAI(
+    api_key="sk-evmlzmwzibqqipnpetyryfxxrbsxeucctkrbppdevuyjvont",
+    base_url="https://api.siliconflow.cn/v1",
+)
+
+# define chat function
+def chat(user_input, max_tokens=11264):
+    # chat history
+    messages_template = [
+  #      {"role": "system", "content": "You are Ling, an assistant created by inclusionAI"},
+        {"role": "user", "content": user_input}
+    ]
+
+    response = client.chat.completions.create(
+        model="inclusionAI/Ling-mini-2.0",
+        messages=messages_template,
+        max_tokens=max_tokens,
+        temperature=0.01,
+        top_p=1,
+        stream=True,
+    )
+    
+    def generate():
+       pass
+
+    resp_text = ""
+    thread = Thread(target=generate)
+    thread.start()
+
+    for chunk in response:
+        if chunk.choices[0].delta.content is not None:
+            resp_text += chunk.choices[0].delta.content
+            yield resp_text
+            print(resp_text)
+
+    thread.join()
+
+
+
+# Create a custom layout using Blocks
+with gr.Blocks(css="""
+    #markdown-output {
+        height: 300px;
+        overflow-y: auto;
+        border: 1px solid #ddd;
+        padding: 10px;
+    }
+""") as demo:
+    gr.Markdown(
+        "## Ling-mini-2.0 AI Assistant\n"
+        "Based on [inclusionAI/Ling-mini-2.0](https://huggingface.co/inclusionAI/Ling-mini-2.0)\n"
+#        "Access through [Ling API](https://bailingchat.alipay.com)"
+    )
+
+    with gr.Row():
+        max_tokens_slider = gr.Slider(minimum=5000, maximum=10000, step=100, label="Generated length")
+
+#    output_box = gr.Textbox(lines=10, label="Response")
+    output_box = gr.Markdown(label="Response", elem_id="markdown-output")
+    input_box = gr.Textbox(lines=8, label="Input you question")
+
+    examples = gr.Examples(
+        examples=[
+            ["Introducing the basic concepts of large language models"],
+            ["How to solve long context dependencies in math problems?"]
+        ],
+        inputs=input_box
+    )
+
+    interface = gr.Interface(
+        fn=chat,
+        inputs=[input_box, max_tokens_slider],
+        outputs=output_box,
+        live=False  # disable auto-triggering on input change
+    )
+
+# launch Gradio Service
+demo.queue()
+demo.launch()
+
+# Construct Gradio Interface
+#interface = gr.Interface(
+#    fn=chat,
+#    inputs=[
+#        gr.Textbox(lines=8, label="输入你的问题"),
+#        gr.Slider(minimum=100, maximum=102400, step=50, label="生成长度")
+#    ],
+#    outputs=[
+#        gr.Textbox(lines=8, label="模型回复")
+#    ],
+#    title="Ling-lite-2.0 AI助手",
+#    description="基于 [inclusionAI/Ling-mini-2.0](https://huggingface.co/inclusionAI/Ling-mini-2.0)  的对话式文本生成演示。",
+#    examples=[
+#        ["介绍大型语言模型的基本概念"],
+#        ["如何解决数学问题中的长上下文依赖？"]
+#    ]
+#)
+
+# launch Gradion Service
+#interface.launch()

requirements.txt

@@ -0,0 +1,5 @@
+gradio
+transformers
+torch
+accelerate
+openai