Spaces:
Runtime error
Runtime error
stage llm model interface
Browse files- README.md +7 -5
- main.py +3 -2
- request_llm/bridge_tgui.py +14 -5
README.md
CHANGED
@@ -36,14 +36,16 @@ https://github.com/polarwinkel/mdtex2html
|
|
36 |
自定义快捷键 | 支持自定义快捷键
|
37 |
配置代理服务器 | 支持配置代理服务器
|
38 |
模块化设计 | 支持自定义高阶的实验性功能
|
39 |
-
自我程序剖析 | [
|
40 |
-
程序剖析 | [
|
41 |
-
读论文 | [
|
42 |
-
|
43 |
-
|
|
|
44 |
公式显示 | 可以同时显示公式的tex形式和渲染形式
|
45 |
图片显示 | 可以在markdown中显示图片
|
46 |
支持GPT输出的markdown表格 | 可以输出支持GPT的markdown表格
|
|
|
47 |
…… | ……
|
48 |
|
49 |
</div>
|
|
|
36 |
自定义快捷键 | 支持自定义快捷键
|
37 |
配置代理服务器 | 支持配置代理服务器
|
38 |
模块化设计 | 支持自定义高阶的实验性功能
|
39 |
+
自我程序剖析 | [函数插件] 一键读懂本项目的源代码
|
40 |
+
程序剖析 | [函数插件] 一键可以剖析其他Python/C++等项目
|
41 |
+
读论文 | [函数插件] 一键解读latex论文全文并生成摘要
|
42 |
+
arxiv小助手 | [函数插件] 输入url一键翻译摘要+下载论文
|
43 |
+
批量注释生成 | [函数插件] 一键批量生成函数注释
|
44 |
+
chat分析报告生成 | [函数插件] 运行后自动生成总结汇报
|
45 |
公式显示 | 可以同时显示公式的tex形式和渲染形式
|
46 |
图片显示 | 可以在markdown中显示图片
|
47 |
支持GPT输出的markdown表格 | 可以输出支持GPT的markdown表格
|
48 |
+
本地大语言模型接口 | 借助[TGUI](https://github.com/oobabooga/text-generation-webui)接入galactica等本地语言模型
|
49 |
…… | ……
|
50 |
|
51 |
</div>
|
main.py
CHANGED
@@ -11,8 +11,9 @@ proxies, WEB_PORT, LLM_MODEL, CONCURRENT_COUNT, AUTHENTICATION, CHATBOT_HEIGHT =
|
|
11 |
PORT = find_free_port() if WEB_PORT <= 0 else WEB_PORT
|
12 |
if not AUTHENTICATION: AUTHENTICATION = None
|
13 |
|
|
|
14 |
initial_prompt = "Serve me as a writing and programming assistant."
|
15 |
-
title_html = "
|
16 |
|
17 |
# 问询记录, python 版本建议3.9+(越新越好)
|
18 |
import logging
|
@@ -140,5 +141,5 @@ def auto_opentab_delay():
|
|
140 |
threading.Thread(target=open, name="open-browser", daemon=True).start()
|
141 |
|
142 |
auto_opentab_delay()
|
143 |
-
demo.title =
|
144 |
demo.queue(concurrency_count=CONCURRENT_COUNT).launch(server_name="0.0.0.0", share=True, server_port=PORT, auth=AUTHENTICATION)
|
|
|
11 |
PORT = find_free_port() if WEB_PORT <= 0 else WEB_PORT
|
12 |
if not AUTHENTICATION: AUTHENTICATION = None
|
13 |
|
14 |
+
title = "ChatGPT 学术优化" if LLM_MODEL.startswith('gpt') else "ChatGPT / LLM 学术优化"
|
15 |
initial_prompt = "Serve me as a writing and programming assistant."
|
16 |
+
title_html = f"<h1 align=\"center\">{title}</h1>"
|
17 |
|
18 |
# 问询记录, python 版本建议3.9+(越新越好)
|
19 |
import logging
|
|
|
141 |
threading.Thread(target=open, name="open-browser", daemon=True).start()
|
142 |
|
143 |
auto_opentab_delay()
|
144 |
+
demo.title = title
|
145 |
demo.queue(concurrency_count=CONCURRENT_COUNT).launch(server_name="0.0.0.0", share=True, server_port=PORT, auth=AUTHENTICATION)
|
request_llm/bridge_tgui.py
CHANGED
@@ -24,9 +24,9 @@ def random_hash():
|
|
24 |
letters = string.ascii_lowercase + string.digits
|
25 |
return ''.join(random.choice(letters) for i in range(9))
|
26 |
|
27 |
-
async def run(context):
|
28 |
params = {
|
29 |
-
'max_new_tokens':
|
30 |
'do_sample': True,
|
31 |
'temperature': 0.5,
|
32 |
'top_p': 0.9,
|
@@ -116,12 +116,15 @@ def predict_tgui(inputs, top_p, temperature, chatbot=[], history=[], system_prom
|
|
116 |
prompt = inputs
|
117 |
tgui_say = ""
|
118 |
|
119 |
-
mutable = [""]
|
120 |
def run_coorotine(mutable):
|
121 |
async def get_result(mutable):
|
122 |
async for response in run(prompt):
|
123 |
print(response[len(mutable[0]):])
|
124 |
mutable[0] = response
|
|
|
|
|
|
|
125 |
asyncio.run(get_result(mutable))
|
126 |
|
127 |
thread_listen = threading.Thread(target=run_coorotine, args=(mutable,), daemon=True)
|
@@ -129,6 +132,7 @@ def predict_tgui(inputs, top_p, temperature, chatbot=[], history=[], system_prom
|
|
129 |
|
130 |
while thread_listen.is_alive():
|
131 |
time.sleep(1)
|
|
|
132 |
# Print intermediate steps
|
133 |
if tgui_say != mutable[0]:
|
134 |
tgui_say = mutable[0]
|
@@ -147,12 +151,17 @@ def predict_tgui_no_ui(inputs, top_p, temperature, history=[], sys_prompt=""):
|
|
147 |
mutable = ["", time.time()]
|
148 |
def run_coorotine(mutable):
|
149 |
async def get_result(mutable):
|
150 |
-
async for response in run(prompt):
|
151 |
print(response[len(mutable[0]):])
|
152 |
mutable[0] = response
|
|
|
|
|
|
|
153 |
asyncio.run(get_result(mutable))
|
154 |
thread_listen = threading.Thread(target=run_coorotine, args=(mutable,))
|
155 |
thread_listen.start()
|
156 |
-
thread_listen.
|
|
|
|
|
157 |
tgui_say = mutable[0]
|
158 |
return tgui_say
|
|
|
24 |
letters = string.ascii_lowercase + string.digits
|
25 |
return ''.join(random.choice(letters) for i in range(9))
|
26 |
|
27 |
+
async def run(context, max_token=512):
|
28 |
params = {
|
29 |
+
'max_new_tokens': max_token,
|
30 |
'do_sample': True,
|
31 |
'temperature': 0.5,
|
32 |
'top_p': 0.9,
|
|
|
116 |
prompt = inputs
|
117 |
tgui_say = ""
|
118 |
|
119 |
+
mutable = ["", time.time()]
|
120 |
def run_coorotine(mutable):
|
121 |
async def get_result(mutable):
|
122 |
async for response in run(prompt):
|
123 |
print(response[len(mutable[0]):])
|
124 |
mutable[0] = response
|
125 |
+
if (time.time() - mutable[1]) > 3:
|
126 |
+
print('exit when no listener')
|
127 |
+
break
|
128 |
asyncio.run(get_result(mutable))
|
129 |
|
130 |
thread_listen = threading.Thread(target=run_coorotine, args=(mutable,), daemon=True)
|
|
|
132 |
|
133 |
while thread_listen.is_alive():
|
134 |
time.sleep(1)
|
135 |
+
mutable[1] = time.time()
|
136 |
# Print intermediate steps
|
137 |
if tgui_say != mutable[0]:
|
138 |
tgui_say = mutable[0]
|
|
|
151 |
mutable = ["", time.time()]
|
152 |
def run_coorotine(mutable):
|
153 |
async def get_result(mutable):
|
154 |
+
async for response in run(prompt, max_token=20):
|
155 |
print(response[len(mutable[0]):])
|
156 |
mutable[0] = response
|
157 |
+
if (time.time() - mutable[1]) > 3:
|
158 |
+
print('exit when no listener')
|
159 |
+
break
|
160 |
asyncio.run(get_result(mutable))
|
161 |
thread_listen = threading.Thread(target=run_coorotine, args=(mutable,))
|
162 |
thread_listen.start()
|
163 |
+
while thread_listen.is_alive():
|
164 |
+
time.sleep(1)
|
165 |
+
mutable[1] = time.time()
|
166 |
tgui_say = mutable[0]
|
167 |
return tgui_say
|