This commit is contained in:
parent
b5036d09c3
commit
2e621b202d
15
meta_ui.py
15
meta_ui.py
|
|
@ -5,15 +5,10 @@ API_URL = "http://localhost:30000/v1/completions"
|
||||||
API_KEY = "token-abc123"
|
API_KEY = "token-abc123"
|
||||||
MODEL_NAME = "Qwen3-14b-base"
|
MODEL_NAME = "Qwen3-14b-base"
|
||||||
|
|
||||||
# 构造 prompt:Base 模型靠拼接上下文
|
# 🚫 不再格式化拼接历史,只取最后一条用户输入
|
||||||
def build_prompt(history, user_message):
|
def build_prompt(history, user_message):
|
||||||
prompt = ""
|
return user_message # 原样作为 prompt 送入模型
|
||||||
for user, bot in history:
|
|
||||||
prompt += f"User: {user}\nAssistant: {bot}\n"
|
|
||||||
prompt += f"User: {user_message}\nAssistant:"
|
|
||||||
return prompt
|
|
||||||
|
|
||||||
# 主对话函数
|
|
||||||
def chat(user_message, history, max_tokens, temperature):
|
def chat(user_message, history, max_tokens, temperature):
|
||||||
prompt = build_prompt(history, user_message)
|
prompt = build_prompt(history, user_message)
|
||||||
|
|
||||||
|
|
@ -25,8 +20,7 @@ def chat(user_message, history, max_tokens, temperature):
|
||||||
"model": MODEL_NAME,
|
"model": MODEL_NAME,
|
||||||
"prompt": prompt,
|
"prompt": prompt,
|
||||||
"max_tokens": max_tokens,
|
"max_tokens": max_tokens,
|
||||||
"temperature": temperature,
|
"temperature": temperature
|
||||||
"stop": ["\nUser:", "\nAssistant:"]
|
|
||||||
}
|
}
|
||||||
|
|
||||||
try:
|
try:
|
||||||
|
|
@ -38,7 +32,6 @@ def chat(user_message, history, max_tokens, temperature):
|
||||||
|
|
||||||
return reply
|
return reply
|
||||||
|
|
||||||
# 手动测试 API 功能
|
|
||||||
def test_api_connection(max_tokens, temperature):
|
def test_api_connection(max_tokens, temperature):
|
||||||
headers = {
|
headers = {
|
||||||
"Authorization": f"Bearer {API_KEY}",
|
"Authorization": f"Bearer {API_KEY}",
|
||||||
|
|
@ -58,7 +51,6 @@ def test_api_connection(max_tokens, temperature):
|
||||||
except Exception as e:
|
except Exception as e:
|
||||||
return f"❌ API 请求失败: {e}"
|
return f"❌ API 请求失败: {e}"
|
||||||
|
|
||||||
# Gradio 控件组合
|
|
||||||
with gr.Blocks(title="Base 模型测试 UI") as demo:
|
with gr.Blocks(title="Base 模型测试 UI") as demo:
|
||||||
gr.Markdown("# 💬 Base 模型对话界面")
|
gr.Markdown("# 💬 Base 模型对话界面")
|
||||||
|
|
||||||
|
|
@ -75,5 +67,4 @@ with gr.Blocks(title="Base 模型测试 UI") as demo:
|
||||||
|
|
||||||
test_btn.click(fn=test_api_connection, inputs=[max_tokens, temperature], outputs=test_output)
|
test_btn.click(fn=test_api_connection, inputs=[max_tokens, temperature], outputs=test_output)
|
||||||
|
|
||||||
# 启动服务
|
|
||||||
demo.launch(server_name="0.0.0.0", server_port=30001)
|
demo.launch(server_name="0.0.0.0", server_port=30001)
|
||||||
|
|
|
||||||
|
|
@ -0,0 +1,79 @@
|
||||||
|
import gradio as gr
|
||||||
|
import requests
|
||||||
|
|
||||||
|
API_URL = "http://localhost:30000/v1/completions"
|
||||||
|
API_KEY = "token-abc123"
|
||||||
|
MODEL_NAME = "Qwen3-14b-base"
|
||||||
|
|
||||||
|
# 构造 prompt:Base 模型靠拼接上下文
|
||||||
|
def build_prompt(history, user_message):
|
||||||
|
prompt = ""
|
||||||
|
for user, bot in history:
|
||||||
|
prompt += f"User: {user}\nAssistant: {bot}\n"
|
||||||
|
prompt += f"User: {user_message}\nAssistant:"
|
||||||
|
return prompt
|
||||||
|
|
||||||
|
# 主对话函数
|
||||||
|
def chat(user_message, history, max_tokens, temperature):
|
||||||
|
prompt = build_prompt(history, user_message)
|
||||||
|
|
||||||
|
headers = {
|
||||||
|
"Authorization": f"Bearer {API_KEY}",
|
||||||
|
"Content-Type": "application/json"
|
||||||
|
}
|
||||||
|
payload = {
|
||||||
|
"model": MODEL_NAME,
|
||||||
|
"prompt": prompt,
|
||||||
|
"max_tokens": max_tokens,
|
||||||
|
"temperature": temperature,
|
||||||
|
"stop": ["\nUser:", "\nAssistant:"]
|
||||||
|
}
|
||||||
|
|
||||||
|
try:
|
||||||
|
response = requests.post(API_URL, headers=headers, json=payload, timeout=30)
|
||||||
|
result = response.json()
|
||||||
|
reply = result["choices"][0]["text"].strip()
|
||||||
|
except Exception as e:
|
||||||
|
reply = f"[请求失败] {e}"
|
||||||
|
|
||||||
|
return reply
|
||||||
|
|
||||||
|
# 手动测试 API 功能
|
||||||
|
def test_api_connection(max_tokens, temperature):
|
||||||
|
headers = {
|
||||||
|
"Authorization": f"Bearer {API_KEY}",
|
||||||
|
"Content-Type": "application/json"
|
||||||
|
}
|
||||||
|
payload = {
|
||||||
|
"model": MODEL_NAME,
|
||||||
|
"prompt": "Ping?",
|
||||||
|
"max_tokens": max_tokens,
|
||||||
|
"temperature": temperature
|
||||||
|
}
|
||||||
|
|
||||||
|
try:
|
||||||
|
resp = requests.post(API_URL, headers=headers, json=payload, timeout=10)
|
||||||
|
out = resp.json()["choices"][0]["text"].strip()
|
||||||
|
return f"✅ API 可用,响应: {out}"
|
||||||
|
except Exception as e:
|
||||||
|
return f"❌ API 请求失败: {e}"
|
||||||
|
|
||||||
|
# Gradio 控件组合
|
||||||
|
with gr.Blocks(title="Base 模型测试 UI") as demo:
|
||||||
|
gr.Markdown("# 💬 Base 模型对话界面")
|
||||||
|
|
||||||
|
with gr.Row():
|
||||||
|
max_tokens = gr.Slider(32, 1024, value=256, label="max_tokens")
|
||||||
|
temperature = gr.Slider(0.0, 1.5, value=0.7, step=0.05, label="temperature")
|
||||||
|
test_btn = gr.Button("🔁 测试 API 可用性")
|
||||||
|
test_output = gr.Textbox(label="API 测试结果", interactive=False)
|
||||||
|
|
||||||
|
chatbot = gr.ChatInterface(
|
||||||
|
fn=lambda msg, hist: chat(msg, hist, max_tokens.value, temperature.value),
|
||||||
|
title=None
|
||||||
|
)
|
||||||
|
|
||||||
|
test_btn.click(fn=test_api_connection, inputs=[max_tokens, temperature], outputs=test_output)
|
||||||
|
|
||||||
|
# 启动服务
|
||||||
|
demo.launch(server_name="0.0.0.0", server_port=30001)
|
||||||
Loading…
Reference in New Issue