Charles Chan
commited on
Commit
·
630d3f4
1
Parent(s):
86b4310
coding
Browse files
app.py
CHANGED
|
@@ -1,9 +1,10 @@
|
|
| 1 |
import streamlit as st
|
|
|
|
| 2 |
from langchain_community.llms import HuggingFaceHub
|
| 3 |
from langchain_community.embeddings import SentenceTransformerEmbeddings
|
| 4 |
from langchain_community.vectorstores import FAISS
|
| 5 |
from datasets import load_dataset
|
| 6 |
-
import
|
| 7 |
|
| 8 |
# 使用 進擊的巨人 数据集
|
| 9 |
try:
|
|
@@ -31,6 +32,7 @@ def answer_question(repo_id, temperature, max_length, question):
|
|
| 31 |
try:
|
| 32 |
with st.spinner("正在初始化 Gemma 模型..."):
|
| 33 |
llm = HuggingFaceHub(repo_id=repo_id, model_kwargs={"temperature": temperature, "max_length": max_length})
|
|
|
|
| 34 |
except Exception as e:
|
| 35 |
st.error(f"Gemma 模型加载失败:{e}")
|
| 36 |
st.stop()
|
|
@@ -49,10 +51,13 @@ def answer_question(repo_id, temperature, max_length, question):
|
|
| 49 |
prompt = f"请根据以下知识库回答问题:\n{context}\n问题:{question}"
|
| 50 |
print('prompt: ' + prompt)
|
| 51 |
|
|
|
|
|
|
|
| 52 |
with st.spinner("正在生成答案..."):
|
| 53 |
answer = llm.invoke(prompt)
|
| 54 |
# 去掉 prompt 的内容
|
| 55 |
answer = answer.replace(prompt, "").strip()
|
|
|
|
| 56 |
return {"prompt": prompt, "answer": answer}
|
| 57 |
except Exception as e:
|
| 58 |
st.error(f"问答过程出错:{e}")
|
|
|
|
| 1 |
import streamlit as st
|
| 2 |
+
import random
|
| 3 |
from langchain_community.llms import HuggingFaceHub
|
| 4 |
from langchain_community.embeddings import SentenceTransformerEmbeddings
|
| 5 |
from langchain_community.vectorstores import FAISS
|
| 6 |
from datasets import load_dataset
|
| 7 |
+
from transformers import pipeline
|
| 8 |
|
| 9 |
# 使用 進擊的巨人 数据集
|
| 10 |
try:
|
|
|
|
| 32 |
try:
|
| 33 |
with st.spinner("正在初始化 Gemma 模型..."):
|
| 34 |
llm = HuggingFaceHub(repo_id=repo_id, model_kwargs={"temperature": temperature, "max_length": max_length})
|
| 35 |
+
st.success("Gemma 模型初始化完成!")
|
| 36 |
except Exception as e:
|
| 37 |
st.error(f"Gemma 模型加载失败:{e}")
|
| 38 |
st.stop()
|
|
|
|
| 51 |
prompt = f"请根据以下知识库回答问题:\n{context}\n问题:{question}"
|
| 52 |
print('prompt: ' + prompt)
|
| 53 |
|
| 54 |
+
st.success("本地数据集筛选完成!")
|
| 55 |
+
|
| 56 |
with st.spinner("正在生成答案..."):
|
| 57 |
answer = llm.invoke(prompt)
|
| 58 |
# 去掉 prompt 的内容
|
| 59 |
answer = answer.replace(prompt, "").strip()
|
| 60 |
+
st.success("答案已经生成!")
|
| 61 |
return {"prompt": prompt, "answer": answer}
|
| 62 |
except Exception as e:
|
| 63 |
st.error(f"问答过程出错:{e}")
|