github
/
ChatGLM-6B
mirror of https://github.com/THUDM/ChatGLM-6B
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

Merge branch 'main' of github.com:THUDM/ChatGLM-6B

pull/477/head
duzx16 2023-04-07 16:44:06 +08:00
parent 0cf46b2348 68488f1f6c
commit 7836739311
2 changed files with 14 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
README.md
View File

@ -35,6 +35,7 @@ ChatGLM-6B 使用了和 ChatGPT 相似的技术，针对中文问答和对话进
* [bibliothecarius](https://github.com/coderabbit214/bibliothecarius)：快速构建服务以集成您的本地数据和AI模型，支持ChatGLM等本地化模型接入。 * [bibliothecarius](https://github.com/coderabbit214/bibliothecarius)：快速构建服务以集成您的本地数据和AI模型，支持ChatGLM等本地化模型接入。
* [闻达](https://github.com/l15y/wenda)：大型语言模型调用平台，基于 ChatGLM-6B 实现了类 ChatPDF 功能 * [闻达](https://github.com/l15y/wenda)：大型语言模型调用平台，基于 ChatGLM-6B 实现了类 ChatPDF 功能
* [JittorLLMs](https://github.com/Jittor/JittorLLMs)：最低3G显存或者没有显卡都可运行 ChatGLM-6B FP16， 支持Linux、windows、Mac部署 * [JittorLLMs](https://github.com/Jittor/JittorLLMs)：最低3G显存或者没有显卡都可运行 ChatGLM-6B FP16， 支持Linux、windows、Mac部署
* [ChatGLM-Finetuning](https://github.com/liucongg/ChatGLM-Finetuning)：基于ChatGLM-6B模型，进行下游具体任务微调，涉及Freeze、Lora、P-tuning等，并进行实验效果对比。
以下是部分针对本项目的教程/文档： 以下是部分针对本项目的教程/文档：
* [Windows部署文档](https://github.com/ZhangErling/ChatGLM-6B/blob/main/deployment_windows.md) * [Windows部署文档](https://github.com/ZhangErling/ChatGLM-6B/blob/main/deployment_windows.md)

16
web_demo2.py
View File

@ -21,7 +21,7 @@ MAX_TURNS = 20
MAX_BOXES = MAX_TURNS * 2 MAX_BOXES = MAX_TURNS * 2
def predict(input, history=None): def predict(input, max_length, top_p, temperature, history=None):
tokenizer, model = get_model() tokenizer, model = get_model()
if history is None: if history is None:
history = [] history = []
@ -35,7 +35,8 @@ def predict(input, history=None):
message(input, avatar_style="big-smile", key=str(len(history)) + "_user") message(input, avatar_style="big-smile", key=str(len(history)) + "_user")
st.write("AI正在回复:") st.write("AI正在回复:")
with st.empty(): with st.empty():
for response, history in model.stream_chat(tokenizer, input, history): for response, history in model.stream_chat(tokenizer, input, history, max_length=max_length, top_p=top_p,
temperature=temperature):
query, response = history[-1] query, response = history[-1]
st.write(response) st.write(response)
@ -49,6 +50,15 @@ prompt_text = st.text_area(label="用户命令输入",
height = 100, height = 100,
placeholder="请在这儿输入您的命令") placeholder="请在这儿输入您的命令")
max_length = st.sidebar.slider(
'max_length', 0, 4096, 2048, step=1
)
top_p = st.sidebar.slider(
'top_p', 0.0, 1.0, 0.6, step=0.01
)
temperature = st.sidebar.slider(
'temperature', 0.0, 1.0, 0.95, step=0.01
)
if 'state' not in st.session_state: if 'state' not in st.session_state:
st.session_state['state'] = [] st.session_state['state'] = []
@ -56,4 +66,4 @@ if 'state' not in st.session_state:
if st.button("发送", key="predict"): if st.button("发送", key="predict"):
with st.spinner("AI正在思考，请稍等........"): with st.spinner("AI正在思考，请稍等........"):
# text generation # text generation
st.session_state["state"] = predict(prompt_text, st.session_state["state"]) st.session_state["state"] = predict(prompt_text, max_length, top_p, temperature, st.session_state["state"])