Spaces:
Sleeping
Sleeping
Update llama3.py
Browse files
llama3.py
CHANGED
|
@@ -1,81 +1,81 @@
|
|
| 1 |
-
import streamlit as st
|
| 2 |
-
import os
|
| 3 |
-
from langchain_groq import ChatGroq
|
| 4 |
-
from langchain_openai import OpenAIEmbeddings
|
| 5 |
-
from langchain.text_splitter import RecursiveCharacterTextSplitter
|
| 6 |
-
from langchain.chains.combine_documents import create_stuff_documents_chain
|
| 7 |
-
from langchain_core.prompts import ChatPromptTemplate
|
| 8 |
-
from langchain.chains import create_retrieval_chain
|
| 9 |
-
from langchain_community.vectorstores import FAISS
|
| 10 |
-
|
| 11 |
-
from langchain_community.document_loaders import PyPDFDirectoryLoader
|
| 12 |
-
|
| 13 |
-
from dotenv import load_dotenv
|
| 14 |
-
|
| 15 |
-
load_dotenv()
|
| 16 |
-
|
| 17 |
-
## load the GroqAPI Key
|
| 18 |
-
os.environ['OPENAI_API_KEY']=os.getenv("OPENAI_API_KEY")
|
| 19 |
-
groq_api_key = os.getenv('GROQ_API_KEY')
|
| 20 |
-
|
| 21 |
-
st.title("ChatBot Demo for Error Codes")
|
| 22 |
-
|
| 23 |
-
llm=ChatGroq(groq_api_key=groq_api_key,
|
| 24 |
-
model="Llama3-8b-8192")
|
| 25 |
-
|
| 26 |
-
|
| 27 |
-
prompt = ChatPromptTemplate.from_template(
|
| 28 |
-
"""
|
| 29 |
-
Answer the question based on the provided context only.
|
| 30 |
-
Please provide the most accurate response based on the question.
|
| 31 |
-
<context>
|
| 32 |
-
{context}
|
| 33 |
-
<context>
|
| 34 |
-
Question: {input}
|
| 35 |
-
"""
|
| 36 |
-
)
|
| 37 |
-
|
| 38 |
-
|
| 39 |
-
def vector_embedding():
|
| 40 |
-
|
| 41 |
-
if "vectors" not in st.session_state:
|
| 42 |
-
|
| 43 |
-
st.session_state.embeddings = OpenAIEmbeddings()
|
| 44 |
-
st.session_state.loader = PyPDFDirectoryLoader("
|
| 45 |
-
st.session_state.docs = st.session_state.loader.load()
|
| 46 |
-
st.session_state.text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
|
| 47 |
-
st.session_state.final_documents = st.session_state.text_splitter.split_documents(st.session_state.docs[:20])
|
| 48 |
-
st.session_state.vectors = FAISS.from_documents(st.session_state.final_documents, st.session_state.embeddings )
|
| 49 |
-
|
| 50 |
-
|
| 51 |
-
|
| 52 |
-
|
| 53 |
-
|
| 54 |
-
prompt1=st.text_input("Enter your question from Documents")
|
| 55 |
-
|
| 56 |
-
if st.button("Documents Embedding"):
|
| 57 |
-
vector_embedding()
|
| 58 |
-
st.write("VectorStore DB is ready")
|
| 59 |
-
|
| 60 |
-
import time
|
| 61 |
-
|
| 62 |
-
|
| 63 |
-
|
| 64 |
-
|
| 65 |
-
if prompt1:
|
| 66 |
-
start = time.process_time()
|
| 67 |
-
document_chain = create_stuff_documents_chain(llm, prompt)
|
| 68 |
-
retriever = st.session_state.vectors.as_retriever()
|
| 69 |
-
retrieval_chain = create_retrieval_chain(retriever, document_chain)
|
| 70 |
-
response = retrieval_chain.invoke({'input': prompt1})
|
| 71 |
-
print("Response time : ", time.process_time() - start)
|
| 72 |
-
st.write(response['answer'])
|
| 73 |
-
|
| 74 |
-
# With a Streamlit expander
|
| 75 |
-
with st.expander("Document Similarity Search"):
|
| 76 |
-
# Find the relevant chunks
|
| 77 |
-
for i, doc in enumerate(response["context"]):
|
| 78 |
-
st.write(doc.page_content)
|
| 79 |
-
st.write("------------------------------------")
|
| 80 |
-
|
| 81 |
-
|
|
|
|
| 1 |
+
import streamlit as st
|
| 2 |
+
import os
|
| 3 |
+
from langchain_groq import ChatGroq
|
| 4 |
+
from langchain_openai import OpenAIEmbeddings
|
| 5 |
+
from langchain.text_splitter import RecursiveCharacterTextSplitter
|
| 6 |
+
from langchain.chains.combine_documents import create_stuff_documents_chain
|
| 7 |
+
from langchain_core.prompts import ChatPromptTemplate
|
| 8 |
+
from langchain.chains import create_retrieval_chain
|
| 9 |
+
from langchain_community.vectorstores import FAISS
|
| 10 |
+
|
| 11 |
+
from langchain_community.document_loaders import PyPDFDirectoryLoader
|
| 12 |
+
|
| 13 |
+
from dotenv import load_dotenv
|
| 14 |
+
|
| 15 |
+
load_dotenv()
|
| 16 |
+
|
| 17 |
+
## load the GroqAPI Key
|
| 18 |
+
os.environ['OPENAI_API_KEY']=os.getenv("OPENAI_API_KEY")
|
| 19 |
+
groq_api_key = os.getenv('GROQ_API_KEY')
|
| 20 |
+
|
| 21 |
+
st.title("ChatBot Demo for Error Codes")
|
| 22 |
+
|
| 23 |
+
llm=ChatGroq(groq_api_key=groq_api_key,
|
| 24 |
+
model="Llama3-8b-8192")
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
prompt = ChatPromptTemplate.from_template(
|
| 28 |
+
"""
|
| 29 |
+
Answer the question based on the provided context only.
|
| 30 |
+
Please provide the most accurate response based on the question.
|
| 31 |
+
<context>
|
| 32 |
+
{context}
|
| 33 |
+
<context>
|
| 34 |
+
Question: {input}
|
| 35 |
+
"""
|
| 36 |
+
)
|
| 37 |
+
|
| 38 |
+
|
| 39 |
+
def vector_embedding():
|
| 40 |
+
|
| 41 |
+
if "vectors" not in st.session_state:
|
| 42 |
+
|
| 43 |
+
st.session_state.embeddings = OpenAIEmbeddings()
|
| 44 |
+
st.session_state.loader = PyPDFDirectoryLoader("/*.pdf")
|
| 45 |
+
st.session_state.docs = st.session_state.loader.load()
|
| 46 |
+
st.session_state.text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
|
| 47 |
+
st.session_state.final_documents = st.session_state.text_splitter.split_documents(st.session_state.docs[:20])
|
| 48 |
+
st.session_state.vectors = FAISS.from_documents(st.session_state.final_documents, st.session_state.embeddings )
|
| 49 |
+
|
| 50 |
+
|
| 51 |
+
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
prompt1=st.text_input("Enter your question from Documents")
|
| 55 |
+
|
| 56 |
+
if st.button("Documents Embedding"):
|
| 57 |
+
vector_embedding()
|
| 58 |
+
st.write("VectorStore DB is ready")
|
| 59 |
+
|
| 60 |
+
import time
|
| 61 |
+
|
| 62 |
+
|
| 63 |
+
|
| 64 |
+
|
| 65 |
+
if prompt1:
|
| 66 |
+
start = time.process_time()
|
| 67 |
+
document_chain = create_stuff_documents_chain(llm, prompt)
|
| 68 |
+
retriever = st.session_state.vectors.as_retriever()
|
| 69 |
+
retrieval_chain = create_retrieval_chain(retriever, document_chain)
|
| 70 |
+
response = retrieval_chain.invoke({'input': prompt1})
|
| 71 |
+
print("Response time : ", time.process_time() - start)
|
| 72 |
+
st.write(response['answer'])
|
| 73 |
+
|
| 74 |
+
# With a Streamlit expander
|
| 75 |
+
with st.expander("Document Similarity Search"):
|
| 76 |
+
# Find the relevant chunks
|
| 77 |
+
for i, doc in enumerate(response["context"]):
|
| 78 |
+
st.write(doc.page_content)
|
| 79 |
+
st.write("------------------------------------")
|
| 80 |
+
|
| 81 |
+
|