Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
@@ -56,7 +56,27 @@ def chatyuan_bot(input, history):
|
|
56 |
history.append((input, output_text))
|
57 |
#print(history)
|
58 |
return history, history
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
59 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
60 |
block = gr.Blocks()
|
61 |
|
62 |
with block as demo:
|
@@ -72,7 +92,10 @@ with block as demo:
|
|
72 |
clear_history = gr.Button("👋 清除历史对话 | Clear")
|
73 |
clear = gr.Button('🧹 清除发送框 | Clear Input')
|
74 |
send = gr.Button("🚀 发送 | Send")
|
75 |
-
|
|
|
|
|
|
|
76 |
send.click(chatyuan_bot, inputs=[message, state], outputs=[chatbot, state])
|
77 |
clear.click(lambda: None, None, message, queue=False)
|
78 |
clear_history.click(fn=clear_session , inputs=[], outputs=[chatbot, state], queue=False)
|
@@ -112,6 +135,8 @@ def chatyuan_bot_api(api_key, input, history):
|
|
112 |
#print(history)
|
113 |
return history, history
|
114 |
|
|
|
|
|
115 |
block = gr.Blocks()
|
116 |
|
117 |
with block as demo_1:
|
@@ -157,23 +182,14 @@ ChatYuan-large-v2是ChatYuan系列中以轻量化实现高质量效果的模型
|
|
157 |
- 增强了模拟情景能力。.<br>
|
158 |
<br>
|
159 |
Based on the original functions of Chatyuan-large-v1, we optimized the model as follows:
|
160 |
-
|
161 |
-Added the ability to speak in both Chinese and English.
|
162 |
-
|
163 |
-Added the ability to refuse to answer. Learn to refuse to answer some dangerous and harmful questions.
|
164 |
-
|
165 |
-Added code generation functionality. Basic code generation has been optimized to a certain extent.
|
166 |
-
|
167 |
-Enhanced basic capabilities. The original contextual Q&A and creative writing skills have significantly improved.
|
168 |
-
|
169 |
-Added a table generation function. Make the generated table content and format more appropriate.
|
170 |
-
|
171 |
-Enhanced basic mathematical computing capabilities.
|
172 |
-
|
173 |
-The maximum number of length tokens has been expanded to 4096.
|
174 |
-
|
175 |
-Enhanced ability to simulate scenarios< br>
|
176 |
-
|
177 |
<br>
|
178 |
👀<a href='https://www.cluebenchmarks.com/clueai.html'>PromptCLUE-large</a>在1000亿token中文语料上预训练, 累计学习1.5万亿中文token, 并且在数百种任务上进行Prompt任务式训练. 针对理解类任务, 如分类、情感分析、抽取等, 可以自定义标签体系; 针对多种生成任务, 可以进行采样自由生成. <br>
|
179 |
<br>
|
|
|
56 |
history.append((input, output_text))
|
57 |
#print(history)
|
58 |
return history, history
|
59 |
+
def chatyuan_bot_regenerate(input, history):
|
60 |
+
|
61 |
+
history = history or []
|
62 |
+
|
63 |
+
if history:
|
64 |
+
input=history[-1][0]
|
65 |
+
history=history[:-1]
|
66 |
+
|
67 |
+
|
68 |
+
if len(history) > 5:
|
69 |
+
history = history[-5:]
|
70 |
|
71 |
+
context = "\n".join([f"用户:{input_text}\n小元:{answer_text}" for input_text, answer_text in history])
|
72 |
+
print(context)
|
73 |
+
|
74 |
+
input_text = context + "\n用户:" + input + "\n小元:"
|
75 |
+
output_text = answer(input_text)
|
76 |
+
history.append((input, output_text))
|
77 |
+
print(history)
|
78 |
+
return history, history
|
79 |
+
|
80 |
block = gr.Blocks()
|
81 |
|
82 |
with block as demo:
|
|
|
92 |
clear_history = gr.Button("👋 清除历史对话 | Clear")
|
93 |
clear = gr.Button('🧹 清除发送框 | Clear Input')
|
94 |
send = gr.Button("🚀 发送 | Send")
|
95 |
+
regenerate = gr.Button("🚀 重新生成本次结果 | regenerate")
|
96 |
+
|
97 |
+
|
98 |
+
regenerate.click(chatyuan_bot_regenerate, inputs=[message, state], outputs=[chatbot, state])
|
99 |
send.click(chatyuan_bot, inputs=[message, state], outputs=[chatbot, state])
|
100 |
clear.click(lambda: None, None, message, queue=False)
|
101 |
clear_history.click(fn=clear_session , inputs=[], outputs=[chatbot, state], queue=False)
|
|
|
135 |
#print(history)
|
136 |
return history, history
|
137 |
|
138 |
+
|
139 |
+
|
140 |
block = gr.Blocks()
|
141 |
|
142 |
with block as demo_1:
|
|
|
182 |
- 增强了模拟情景能力。.<br>
|
183 |
<br>
|
184 |
Based on the original functions of Chatyuan-large-v1, we optimized the model as follows:
|
|
|
185 |
-Added the ability to speak in both Chinese and English.
|
|
|
186 |
-Added the ability to refuse to answer. Learn to refuse to answer some dangerous and harmful questions.
|
|
|
187 |
-Added code generation functionality. Basic code generation has been optimized to a certain extent.
|
|
|
188 |
-Enhanced basic capabilities. The original contextual Q&A and creative writing skills have significantly improved.
|
|
|
189 |
-Added a table generation function. Make the generated table content and format more appropriate.
|
|
|
190 |
-Enhanced basic mathematical computing capabilities.
|
|
|
191 |
-The maximum number of length tokens has been expanded to 4096.
|
|
|
192 |
-Enhanced ability to simulate scenarios< br>
|
|
|
193 |
<br>
|
194 |
👀<a href='https://www.cluebenchmarks.com/clueai.html'>PromptCLUE-large</a>在1000亿token中文语料上预训练, 累计学习1.5万亿中文token, 并且在数百种任务上进行Prompt任务式训练. 针对理解类任务, 如分类、情感分析、抽取等, 可以自定义标签体系; 针对多种生成任务, 可以进行采样自由生成. <br>
|
195 |
<br>
|