Upload 3 files

lawmain.py

@@ -1,6 +1,6 @@
 import streamlit as st
 from PIL import Image
-from lawchain import get_lpphelper_chain,process_llm_response
+from lppchain import get_lpphelper_chain,process_llm_response
 
 #st.title( "Lakna Reddy & Associates 🤖")
 col1, mid, col2 = st.columns(3)

lawtrainmodel.py

@@ -0,0 +1,50 @@
+import transformers
+import torch
+import os
+from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
+from transformers import pipeline
+from langchain.llms import HuggingFacePipeline
+from langchain.vectorstores import Chroma
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+from langchain.chains import RetrievalQA
+from langchain.document_loaders import TextLoader
+from langchain.document_loaders import PyPDFLoader
+from langchain.document_loaders import DirectoryLoader
+from InstructorEmbedding import INSTRUCTOR
+from langchain.embeddings import HuggingFaceInstructEmbeddings
+from langchain_community.vectorstores import Chroma
+import textwrap
+
+def gen_vectordb():
+    tokenizer = AutoTokenizer.from_pretrained("lmsys/fastchat-t5-3b-v1.0")
+    model = AutoModelForSeq2SeqLM.from_pretrained("lmsys/fastchat-t5-3b-v1.0")
+    pipe = pipeline(
+        "text2text-generation",
+        model=model, 
+        tokenizer=tokenizer, 
+        max_length=256
+    )
+
+    local_llm = HuggingFacePipeline(pipeline=pipe)
+    loader = DirectoryLoader('C:/Users/SudheerRChinthala/sivallm/new_papers', glob="./*.pdf", loader_cls=PyPDFLoader)
+    documents = loader.load()
+    text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
+    texts = text_splitter.split_documents(documents)
+
+    instructor_embeddings = HuggingFaceInstructEmbeddings(model_name="hkunlp/instructor-base") 
+    persist_directory = 'db'
+    embedding = instructor_embeddings
+    vectordb = Chroma.from_documents(documents=texts, 
+                                    embedding=embedding,
+                                    persist_directory=persist_directory)
+    retriever = vectordb.as_retriever(search_kwargs={"k": 3})
+    qa_chain = RetrievalQA.from_chain_type(llm=local_llm, 
+                                    chain_type="stuff", 
+                                    retriever=retriever, 
+                                    return_source_documents=True)
+    vectordb.persist()
+    vectordb = None 
+
+
+if __name__=="__main__":
+    gen_vectordb()