Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
@@ -1,132 +1,247 @@
|
|
1 |
import gradio as gr
|
|
|
|
|
2 |
from huggingface_hub import InferenceClient
|
3 |
-
from typing import List, Tuple
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
4 |
|
5 |
-
# Default settings
|
6 |
class ChatConfig:
|
|
|
7 |
MODEL = "google/gemma-3-27b-it"
|
8 |
DEFAULT_SYSTEM_MSG = "You are a super intelligent and useful Chatbot."
|
9 |
DEFAULT_MAX_TOKENS = 512
|
10 |
DEFAULT_TEMP = 0.3
|
11 |
DEFAULT_TOP_P = 0.95
|
|
|
12 |
|
13 |
-
|
14 |
-
|
15 |
-
def generate_response(
|
16 |
-
message: str,
|
17 |
-
history: List[Tuple[str, str]],
|
18 |
-
system_message: str = ChatConfig.DEFAULT_SYSTEM_MSG,
|
19 |
-
max_tokens: int = ChatConfig.DEFAULT_MAX_TOKENS,
|
20 |
-
temperature: float = ChatConfig.DEFAULT_TEMP,
|
21 |
-
top_p: float = ChatConfig.DEFAULT_TOP_P
|
22 |
-
) -> str:
|
23 |
-
messages = [{"role": "system", "content": system_message}]
|
24 |
|
25 |
-
|
26 |
-
|
27 |
-
|
28 |
-
|
29 |
-
|
30 |
-
|
|
|
|
|
|
|
|
|
|
|
31 |
|
32 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
33 |
|
34 |
-
|
35 |
-
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
|
41 |
-
|
42 |
-
|
43 |
-
|
44 |
-
|
45 |
-
|
46 |
-
|
47 |
-
|
48 |
-
|
49 |
-
|
50 |
-
|
51 |
-
|
52 |
-
|
53 |
-
|
54 |
-
|
55 |
-
|
56 |
-
|
57 |
-
|
58 |
-
|
59 |
-
|
60 |
-
|
61 |
-
|
62 |
-
|
63 |
-
|
64 |
-
|
65 |
-
|
66 |
-
|
67 |
-
|
68 |
-
|
69 |
-
|
70 |
-
|
71 |
-
|
72 |
-
|
73 |
-
|
74 |
-
|
75 |
-
|
76 |
-
|
77 |
-
|
78 |
-
|
79 |
-
|
80 |
-
|
81 |
-
|
82 |
-
|
83 |
-
|
84 |
-
|
85 |
-
|
86 |
-
|
87 |
-
|
88 |
-
|
89 |
-
|
90 |
-
|
91 |
-
|
92 |
-
|
93 |
-
|
94 |
-
|
95 |
-
|
96 |
-
|
97 |
-
|
98 |
-
|
99 |
-
|
100 |
-
|
101 |
-
|
102 |
-
|
103 |
-
|
104 |
-
|
105 |
-
|
106 |
-
|
107 |
-
|
108 |
-
|
109 |
-
|
110 |
-
|
111 |
-
|
112 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
113 |
)
|
114 |
-
|
115 |
-
|
116 |
-
|
117 |
-
|
118 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
119 |
|
120 |
def main():
|
121 |
-
|
122 |
-
|
123 |
-
|
124 |
-
|
125 |
-
|
126 |
-
|
127 |
-
|
128 |
-
|
129 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
130 |
|
131 |
if __name__ == "__main__":
|
132 |
main()
|
|
|
1 |
import gradio as gr
|
2 |
+
import os
|
3 |
+
import logging
|
4 |
from huggingface_hub import InferenceClient
|
5 |
+
from typing import List, Tuple, Generator, Dict, Any, Optional
|
6 |
+
|
7 |
+
# Configure logging
|
8 |
+
logging.basicConfig(
|
9 |
+
level=logging.INFO,
|
10 |
+
format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
|
11 |
+
)
|
12 |
+
logger = logging.getLogger(__name__)
|
13 |
|
|
|
14 |
class ChatConfig:
|
15 |
+
"""Configuration settings for the chat application."""
|
16 |
MODEL = "google/gemma-3-27b-it"
|
17 |
DEFAULT_SYSTEM_MSG = "You are a super intelligent and useful Chatbot."
|
18 |
DEFAULT_MAX_TOKENS = 512
|
19 |
DEFAULT_TEMP = 0.3
|
20 |
DEFAULT_TOP_P = 0.95
|
21 |
+
HF_TOKEN = os.environ.get("HF_TOKEN", None) # Get token from environment variable
|
22 |
|
23 |
+
class ChatApp:
|
24 |
+
"""Main chat application class."""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
25 |
|
26 |
+
def __init__(self):
|
27 |
+
"""Initialize the chat application."""
|
28 |
+
try:
|
29 |
+
self.client = InferenceClient(
|
30 |
+
ChatConfig.MODEL,
|
31 |
+
token=ChatConfig.HF_TOKEN
|
32 |
+
)
|
33 |
+
logger.info(f"Successfully initialized InferenceClient for {ChatConfig.MODEL}")
|
34 |
+
except Exception as e:
|
35 |
+
logger.error(f"Failed to initialize InferenceClient: {e}")
|
36 |
+
raise
|
37 |
|
38 |
+
def generate_response(
|
39 |
+
self,
|
40 |
+
message: str,
|
41 |
+
history: List[Tuple[str, str]],
|
42 |
+
system_message: str = ChatConfig.DEFAULT_SYSTEM_MSG,
|
43 |
+
max_tokens: int = ChatConfig.DEFAULT_MAX_TOKENS,
|
44 |
+
temperature: float = ChatConfig.DEFAULT_TEMP,
|
45 |
+
top_p: float = ChatConfig.DEFAULT_TOP_P
|
46 |
+
) -> Generator[str, None, None]:
|
47 |
+
"""Generate streaming responses from the model."""
|
48 |
+
if not message.strip():
|
49 |
+
yield "请输入消息。"
|
50 |
+
return
|
51 |
+
|
52 |
+
messages = [{"role": "system", "content": system_message}]
|
53 |
+
|
54 |
+
# Add conversation history
|
55 |
+
for user_msg, bot_msg in history:
|
56 |
+
if user_msg:
|
57 |
+
messages.append({"role": "user", "content": user_msg})
|
58 |
+
if bot_msg:
|
59 |
+
messages.append({"role": "assistant", "content": bot_msg})
|
60 |
+
|
61 |
+
# Add the current message
|
62 |
+
messages.append({"role": "user", "content": message})
|
63 |
+
|
64 |
+
try:
|
65 |
+
response = ""
|
66 |
+
for chunk in self.client.chat_completion(
|
67 |
+
messages,
|
68 |
+
max_tokens=max_tokens,
|
69 |
+
stream=True,
|
70 |
+
temperature=temperature,
|
71 |
+
top_p=top_p,
|
72 |
+
):
|
73 |
+
token = chunk.choices[0].delta.content or ""
|
74 |
+
response += token
|
75 |
+
yield response
|
76 |
+
except Exception as e:
|
77 |
+
logger.error(f"Error generating response: {e}")
|
78 |
+
yield f"抱歉,发生了错误: {str(e)}"
|
79 |
|
80 |
+
def create_interface(self) -> gr.Blocks:
|
81 |
+
"""Create and configure the chat interface."""
|
82 |
+
# Custom CSS for a modern look
|
83 |
+
custom_css = """
|
84 |
+
.chatbot .message {
|
85 |
+
border-radius: 12px;
|
86 |
+
margin: 8px;
|
87 |
+
padding: 12px;
|
88 |
+
box-shadow: 0 1px 3px rgba(0,0,0,0.12);
|
89 |
+
}
|
90 |
+
.chatbot .user-message {
|
91 |
+
background-color: #e3f2fd;
|
92 |
+
border-left: 4px solid #2196F3;
|
93 |
+
}
|
94 |
+
.chatbot .bot-message {
|
95 |
+
background-color: #f5f5f5;
|
96 |
+
border-left: 4px solid #9e9e9e;
|
97 |
+
}
|
98 |
+
.gr-button {
|
99 |
+
border-radius: 8px;
|
100 |
+
padding: 8px 16px;
|
101 |
+
transition: all 0.3s ease;
|
102 |
+
}
|
103 |
+
.gr-button:hover {
|
104 |
+
transform: translateY(-2px);
|
105 |
+
box-shadow: 0 4px 8px rgba(0,0,0,0.15);
|
106 |
+
}
|
107 |
+
.container {
|
108 |
+
max-width: 900px;
|
109 |
+
margin: 0 auto;
|
110 |
+
}
|
111 |
+
"""
|
112 |
+
|
113 |
+
with gr.Blocks(css=custom_css, theme=gr.themes.Soft()) as interface:
|
114 |
+
gr.Markdown("# 喵哥Google-Gemma-3尝鲜版")
|
115 |
+
gr.Markdown("与Google Gemma 3 27B模型互动,可自定义参数。")
|
116 |
+
|
117 |
+
with gr.Row():
|
118 |
+
with gr.Column(scale=4):
|
119 |
+
chatbot = gr.Chatbot(
|
120 |
+
label="Gemma Chat",
|
121 |
+
avatar_images=("./user.png", "./botge.png"),
|
122 |
+
height=500,
|
123 |
+
show_copy_button=True,
|
124 |
+
elem_classes="chatbox"
|
125 |
+
)
|
126 |
+
|
127 |
+
with gr.Row():
|
128 |
+
with gr.Column(scale=8):
|
129 |
+
msg = gr.Textbox(
|
130 |
+
show_label=False,
|
131 |
+
placeholder="在这里输入您的消息...",
|
132 |
+
container=False,
|
133 |
+
lines=2
|
134 |
+
)
|
135 |
+
with gr.Column(scale=1, min_width=70):
|
136 |
+
submit_btn = gr.Button("发送", variant="primary")
|
137 |
+
|
138 |
+
with gr.Row():
|
139 |
+
clear_btn = gr.Button("清空对话", variant="secondary")
|
140 |
+
example_btn = gr.Button("加载示例", variant="secondary")
|
141 |
+
|
142 |
+
with gr.Column(scale=2):
|
143 |
+
with gr.Accordion("聊天设置", open=True):
|
144 |
+
system_msg = gr.Textbox(
|
145 |
+
value=ChatConfig.DEFAULT_SYSTEM_MSG,
|
146 |
+
label="系统提示词",
|
147 |
+
lines=3,
|
148 |
+
placeholder="输入系统提示词..."
|
149 |
+
)
|
150 |
+
|
151 |
+
with gr.Accordion("高级参数", open=False):
|
152 |
+
max_tokens = gr.Slider(
|
153 |
+
minimum=1,
|
154 |
+
maximum=8192,
|
155 |
+
value=ChatConfig.DEFAULT_MAX_TOKENS,
|
156 |
+
step=1,
|
157 |
+
label="最大标记数",
|
158 |
+
info="控制回复长度"
|
159 |
+
)
|
160 |
+
temperature = gr.Slider(
|
161 |
+
minimum=0.1,
|
162 |
+
maximum=1.0,
|
163 |
+
value=ChatConfig.DEFAULT_TEMP,
|
164 |
+
step=0.1,
|
165 |
+
label="温度",
|
166 |
+
info="控制随机性"
|
167 |
+
)
|
168 |
+
top_p = gr.Slider(
|
169 |
+
minimum=0.1,
|
170 |
+
maximum=1.0,
|
171 |
+
value=ChatConfig.DEFAULT_TOP_P,
|
172 |
+
step=0.05,
|
173 |
+
label="Top-P",
|
174 |
+
info="控制多样性"
|
175 |
+
)
|
176 |
+
|
177 |
+
with gr.Accordion("示例问题", open=False):
|
178 |
+
gr.Examples(
|
179 |
+
examples=[
|
180 |
+
["讲一讲人工智能的最新进展"],
|
181 |
+
["写一个关于机器���发现情感的短篇故事"],
|
182 |
+
["向10岁的孩子解释量子计算"]
|
183 |
+
],
|
184 |
+
inputs=msg
|
185 |
+
)
|
186 |
+
|
187 |
+
with gr.Accordion("模型信息", open=False):
|
188 |
+
gr.Markdown(f"""
|
189 |
+
- **模型**: {ChatConfig.MODEL}
|
190 |
+
- **提供商**: Hugging Face
|
191 |
+
- **描述**: Gemma 3是Google推出的先进开源大语言模型。
|
192 |
+
""")
|
193 |
+
|
194 |
+
# Set up event handlers
|
195 |
+
msg_and_submit = [msg, submit_btn]
|
196 |
+
|
197 |
+
submit_click = submit_btn.click(
|
198 |
+
fn=self.generate_response,
|
199 |
+
inputs=[msg, chatbot, system_msg, max_tokens, temperature, top_p],
|
200 |
+
outputs=chatbot,
|
201 |
+
api_name="chat"
|
202 |
)
|
203 |
+
|
204 |
+
# Submit when pressing Enter (but not when pressing Shift+Enter)
|
205 |
+
msg.submit(
|
206 |
+
fn=self.generate_response,
|
207 |
+
inputs=[msg, chatbot, system_msg, max_tokens, temperature, top_p],
|
208 |
+
outputs=chatbot,
|
209 |
+
api_name=False
|
210 |
+
)
|
211 |
+
|
212 |
+
# Clear the textbox after sending
|
213 |
+
submit_click.then(lambda: "", None, msg)
|
214 |
+
msg.submit(lambda: "", None, msg)
|
215 |
+
|
216 |
+
# Clear chat button
|
217 |
+
clear_btn.click(lambda: None, None, chatbot)
|
218 |
+
|
219 |
+
# Example button
|
220 |
+
example_btn.click(
|
221 |
+
lambda: ("介绍一下人工智能研究中最有趣的发展", []),
|
222 |
+
None,
|
223 |
+
[msg, chatbot]
|
224 |
+
)
|
225 |
+
|
226 |
+
return interface
|
227 |
|
228 |
def main():
|
229 |
+
"""Main function to launch the application."""
|
230 |
+
try:
|
231 |
+
app = ChatApp()
|
232 |
+
interface = app.create_interface()
|
233 |
+
|
234 |
+
interface.launch(
|
235 |
+
server_name="0.0.0.0",
|
236 |
+
server_port=7860,
|
237 |
+
share=False,
|
238 |
+
show_api=False,
|
239 |
+
debug=True,
|
240 |
+
show_error=True
|
241 |
+
)
|
242 |
+
except Exception as e:
|
243 |
+
logger.critical(f"Failed to launch application: {e}")
|
244 |
+
print(f"错误: {e}")
|
245 |
|
246 |
if __name__ == "__main__":
|
247 |
main()
|