starlight-apk/Qwen
1
0
Fork 0
mirror of https://github.com/QwenLM/Qwen.git synced 2026-05-20 08:25:47 +08:00
Code Issues Packages Projects Releases Wiki Activity

update web_demo

Browse Source
This commit is contained in:
yangapku
2023-08-10 21:47:23 +08:00
parent 867879545f
commit 2e5cff6454
5 changed files with 123 additions and 94 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
README.md
View File

@@ -257,7 +257,7 @@ python cli_demo.py
We provide code for users to build a web UI demo (thanks to @wysiad). Before you start, make sure you install the following packages: We provide code for users to build a web UI demo (thanks to @wysiad). Before you start, make sure you install the following packages:
``` ```
pip install gradio mdtex2html pip install -r requirements_web_demo.txt
``` ```
Then run the command below and click on the generated link: Then run the command below and click on the generated link:

2
README_CN.md
View File

@@ -259,7 +259,7 @@ python cli_demo.py
我们提供了Web UI的demo供用户使用 (感谢 @wysiad 支持)。在开始前，确保已经安装如下代码库： 我们提供了Web UI的demo供用户使用 (感谢 @wysiad 支持)。在开始前，确保已经安装如下代码库：
``` ```
pip install gradio mdtex2html pip install -r requirements_web_demo.txt
``` ```
随后运行如下命令，并点击生成链接： 随后运行如下命令，并点击生成链接：

2
README_JA.md
View File

@@ -264,7 +264,7 @@ python cli_demo.py
ウェブUIデモを構築するためのコードを提供します（@wysiadに感謝）。始める前に、以下のパッケージがインストールされていることを確認してください： ウェブUIデモを構築するためのコードを提供します（@wysiadに感謝）。始める前に、以下のパッケージがインストールされていることを確認してください：
``` ```
pip install gradio mdtex2html pip install -r requirements_web_demo.txt
``` ```
そして、以下のコマンドを実行し、生成されたリンクをクリックする： そして、以下のコマンドを実行し、生成されたリンクをクリックする：

8
cli_demo.py
View File

@@ -38,7 +38,7 @@ Commands:
def _load_model_tokenizer(args): def _load_model_tokenizer(args):
tokenizer = AutoTokenizer.from_pretrained( tokenizer = AutoTokenizer.from_pretrained(
args.checkpoint_path, trust_remote_code=True, args.checkpoint_path, trust_remote_code=True, resume_download=True,
) )
if args.cpu_only: if args.cpu_only:
@@ -50,9 +50,11 @@ def _load_model_tokenizer(args):
args.checkpoint_path, args.checkpoint_path,
device_map=device_map, device_map=device_map,
trust_remote_code=True, trust_remote_code=True,
resume_download=True,
).eval() ).eval()
model.generation_config = GenerationConfig.from_pretrained(args.checkpoint_path, trust_remote_code=True) model.generation_config = GenerationConfig.from_pretrained(
args.checkpoint_path, trust_remote_code=True, resume_download=True,
)
return model, tokenizer return model, tokenizer

195
web_demo.py
View File

@@ -1,52 +1,60 @@
#!/usr/bin/env python3 # Copyright (c) Alibaba Cloud.
#
# This source code is licensed under the license found in the
# LICENSE file in the root directory of this source tree.
"""A simple web interactive chat demo based on gradio."""
from argparse import ArgumentParser
from transformers import AutoTokenizer
import gradio as gr import gradio as gr
import mdtex2html import mdtex2html
from transformers import AutoModelForCausalLM, AutoTokenizer from transformers import AutoModelForCausalLM, AutoTokenizer
from transformers.generation import GenerationConfig from transformers.generation import GenerationConfig
from argparse import ArgumentParser
import sys
print("Call args:" + str(sys.argv)) DEFAULT_CKPT_PATH = 'QWen/QWen-7B-Chat'
parser = ArgumentParser()
parser.add_argument("--share", action="store_true", default=False)
parser.add_argument("--inbrowser", action="store_true", default=False)
parser.add_argument("--server_port", type=int, default=80)
parser.add_argument("--server_name", type=str, default="0.0.0.0")
parser.add_argument("--exit", action="store_true", default=False)
parser.add_argument("--model_revision", type=str, default="")
args = parser.parse_args(sys.argv[1:])
print("Args:" + str(args))
tokenizer = AutoTokenizer.from_pretrained(
"Qwen/Qwen-7B-Chat", trust_remote_code=True, resume_download=True
)
model = AutoModelForCausalLM.from_pretrained( def _get_args():
"Qwen/Qwen-7B-Chat", parser = ArgumentParser()
device_map="auto", parser.add_argument("-c", "--checkpoint-path", type=str, default=DEFAULT_CKPT_PATH,
help="Checkpoint name or path, default to %(default)r")
parser.add_argument("--cpu-only", action="store_true", help="Run demo with CPU only")
parser.add_argument("--share", action="store_true", default=False,
help="Create a publicly shareable link for the interface.")
parser.add_argument("--inbrowser", action="store_true", default=False,
help="Automatically launch the interface in a new tab on the default browser.")
parser.add_argument("--server-port", type=int, default=8000,
help="Demo server port.")
parser.add_argument("--server-name", type=str, default="127.0.0.1",
help="Demo server name.")
args = parser.parse_args()
return args
def _load_model_tokenizer(args):
tokenizer = AutoTokenizer.from_pretrained(
args.checkpoint_path, trust_remote_code=True, resume_download=True,
)
if args.cpu_only:
device_map = "cpu"
else:
device_map = "auto"
model = AutoModelForCausalLM.from_pretrained(
args.checkpoint_path,
device_map=device_map,
trust_remote_code=True, trust_remote_code=True,
resume_download=True, resume_download=True,
**{"revision": args.model_revision} ).eval()
if args.model_revision is not None model.generation_config = GenerationConfig.from_pretrained(
and args.model_revision != "" args.checkpoint_path, trust_remote_code=True, resume_download=True,
and args.model_revision != "None" )
else {},
).eval()
model.generation_config = GenerationConfig.from_pretrained( return model, tokenizer
"Qwen/Qwen-7B-Chat", trust_remote_code=True, resume_download=True
)
if "exit" in args:
if args.exit:
sys.exit(0)
else:
del args.exit
if "model_revision" in args:
del args.model_revision
def postprocess(self, y): def postprocess(self, y):
@@ -54,7 +62,7 @@ def postprocess(self, y):
return [] return []
for i, (message, response) in enumerate(y): for i, (message, response) in enumerate(y):
y[i] = ( y[i] = (
None if message is None else mdtex2html.convert((message)), None if message is None else mdtex2html.convert(message),
None if response is None else mdtex2html.convert(response), None if response is None else mdtex2html.convert(response),
) )
return y return y
@@ -63,7 +71,7 @@ def postprocess(self, y):
gr.Chatbot.postprocess = postprocess gr.Chatbot.postprocess = postprocess
def parse_text(text): def _parse_text(text):
lines = text.split("\n") lines = text.split("\n")
lines = [line for line in lines if line != ""] lines = [line for line in lines if line != ""]
count = 0 count = 0
@@ -78,7 +86,7 @@ def parse_text(text):
else: else:
if i > 0: if i > 0:
if count % 2 == 1: if count % 2 == 1:
line = line.replace("`", "\`") line = line.replace("`", r"\`")
line = line.replace("<", "&lt;") line = line.replace("<", "&lt;")
line = line.replace(">", "&gt;") line = line.replace(">", "&gt;")
line = line.replace(" ", "&nbsp;") line = line.replace(" ", "&nbsp;")
@@ -95,70 +103,89 @@ def parse_text(text):
return text return text
task_history = [] def _launch_demo(args, model, tokenizer):
task_history = []
def predict(_query, _chatbot):
print("User: " + _parse_text(_query))
_chatbot.append((_parse_text(_query), ""))
full_response = ""
def predict(query, chatbot): for response in model.chat_stream(tokenizer, _query, history=task_history):
print("User: " + parse_text(query)) _chatbot[-1] = (_parse_text(_query), _parse_text(response))
chatbot.append((parse_text(query), ""))
fullResponse = ""
for response in model.chat_stream(tokenizer, query, history=task_history): yield _chatbot
chatbot[-1] = (parse_text(query), parse_text(response)) full_response = _parse_text(response)
yield chatbot task_history.append((_query, full_response))
fullResponse = parse_text(response) print("Qwen-7B-Chat: " + _parse_text(full_response))
task_history.append((query, fullResponse)) def regenerate(_chatbot):
print("Qwen-7B-Chat: " + parse_text(fullResponse))
def regenerate(chatbot):
if not task_history: if not task_history:
yield chatbot yield _chatbot
return return
item = task_history.pop(-1) item = task_history.pop(-1)
chatbot.pop(-1) _chatbot.pop(-1)
yield from predict(item[0], chatbot) yield from predict(item[0], _chatbot)
def reset_user_input():
def reset_user_input():
return gr.update(value="") return gr.update(value="")
def reset_state():
def reset_state():
task_history.clear() task_history.clear()
return [] return []
with gr.Blocks() as demo:
with gr.Blocks() as demo: gr.Markdown("""\
gr.Markdown("""<p align="center"><img src="https://modelscope.cn/api/v1/models/qwen/Qwen-7B-Chat/repo?Revision=master&FilePath=assets/logo.jpeg&View=true" style="height: 80px"/><p>""") <p align="center"><img src="https://modelscope.cn/api/v1/models/qwen/Qwen-7B-Chat/repo?
Revision=master&FilePath=assets/logo.jpeg&View=true" style="height: 80px"/><p>""")
gr.Markdown("""<center><font size=8>Qwen-7B-Chat Bot</center>""") gr.Markdown("""<center><font size=8>Qwen-7B-Chat Bot</center>""")
gr.Markdown( gr.Markdown(
"""<center><font size=3>This WebUI is based on Qwen-7B-Chat, developed by Alibaba Cloud. (本WebUI基于Qwen-7B-Chat打造，实现聊天机器人功能。)</center>""" """\
) <center><font size=3>This WebUI is based on Qwen-7B-Chat, developed by Alibaba Cloud. \
gr.Markdown( (本WebUI基于Qwen-7B-Chat打造，实现聊天机器人功能。)</center>""")
"""<center><font size=4>Qwen-7B <a href="https://modelscope.cn/models/qwen/Qwen-7B/summary">🤖 <a> | <a href="https://huggingface.co/Qwen/Qwen-7B">🤗</a>&nbsp ｜ Qwen-7B-Chat <a href="https://modelscope.cn/models/qwen/Qwen-7B-Chat/summary">🤖 <a>| <a href="https://huggingface.co/Qwen/Qwen-7B-Chat">🤗</a>&nbsp ｜ &nbsp<a href="https://github.com/QwenLM/Qwen-7B">Github</a></center>""" gr.Markdown("""\
) <center><font size=4>Qwen-7B <a href="https://modelscope.cn/models/qwen/Qwen-7B/summary">🤖 </a>
| <a href="https://huggingface.co/Qwen/Qwen-7B">🤗</a>&nbsp ｜
Qwen-7B-Chat <a href="https://modelscope.cn/models/qwen/Qwen-7B-Chat/summary">🤖 </a> |
<a href="https://huggingface.co/Qwen/Qwen-7B-Chat">🤗</a>&nbsp ｜
&nbsp<a href="https://github.com/QwenLM/Qwen-7B">Github</a></center>""")
chatbot = gr.Chatbot(lines=10, label='Qwen-7B-Chat', elem_classes="control-height") chatbot = gr.Chatbot(label='Qwen-7B-Chat', elem_classes="control-height")
query = gr.Textbox(lines=2, label='Input') query = gr.Textbox(lines=2, label='Input')
with gr.Row(): with gr.Row():
emptyBtn = gr.Button("🧹 Clear History (清除历史)") empty_btn = gr.Button("🧹 Clear History (清除历史)")
submitBtn = gr.Button("🚀 Submit (发送)") submit_btn = gr.Button("🚀 Submit (发送)")
regenBtn = gr.Button("🤔️ Regenerate (重试)") regen_btn = gr.Button("🤔️ Regenerate (重试)")
submitBtn.click(predict, [query, chatbot], [chatbot], show_progress=True) submit_btn.click(predict, [query, chatbot], [chatbot], show_progress=True)
submitBtn.click(reset_user_input, [], [query]) submit_btn.click(reset_user_input, [], [query])
emptyBtn.click(reset_state, outputs=[chatbot], show_progress=True) empty_btn.click(reset_state, outputs=[chatbot], show_progress=True)
regenBtn.click(regenerate, [chatbot], [chatbot], show_progress=True) regen_btn.click(regenerate, [chatbot], [chatbot], show_progress=True)
gr.Markdown( gr.Markdown("""\
"""<font size=2>Note: This demo is governed by the original license of Qwen-7B. We strongly advise users not to knowingly generate or allow others to knowingly generate harmful content, including hate speech, violence, pornography, deception, etc. (注：本演示受Qwen-7B的许可协议限制。我们强烈建议，用户不应传播及不应允许他人传播以下内容，包括但不限于仇恨言论、暴力、色情、欺诈相关的有害信息。)""" <font size=2>Note: This demo is governed by the original license of Qwen-7B. \
We strongly advise users not to knowingly generate or allow others to knowingly generate harmful content, \
including hate speech, violence, pornography, deception, etc. \
(注：本演示受Qwen-7B的许可协议限制。我们强烈建议，用户不应传播及不应允许他人传播以下内容，\
包括但不限于仇恨言论、暴力、色情、欺诈相关的有害信息。)""")
demo.queue().launch(
share=args.share,
inbrowser=args.inbrowser,
server_port=args.server_port,
server_name=args.server_name,
) )
if len(sys.argv) > 1:
demo.queue().launch(**vars(args)) def main():
else: args = _get_args()
demo.queue().launch()
model, tokenizer = _load_model_tokenizer(args)
_launch_demo(args, model, tokenizer)
if __name__ == '__main__':
main()