Update app.py

app.py

@@ -12,8 +12,8 @@ from langchain_core.prompts import ChatPromptTemplate
 import os
 from dotenv import load_dotenv
 from helper import SYSTEM_PROMPT
-from langchain_google_genai import GoogleGenerativeAIEmbeddings
-# from langchain.embeddings import HuggingFaceEmbeddings # open source free embedding
+# from langchain_google_genai import GoogleGenerativeAIEmbeddings
+from langchain.embeddings import HuggingFaceEmbeddings # open source free embedding
 load_dotenv()
 
 
@@ -23,9 +23,9 @@ class PDFQAProcessor:
 
     llm = ChatGroq(
         # model_name="deepseek-r1-distill-llama-70b",
-        model_name="llama3-70b-8192",
+        model_name="llama-3.3-70b-versatile",
         temperature=0.1,
-        max_tokens=3000,
+        max_tokens=8192,
         api_key = os.getenv('GROQ_API_KEY')
     )
 
@@ -37,15 +37,15 @@ class PDFQAProcessor:
             
     question_answer_chain = create_stuff_documents_chain(llm, prompt)
 
-    # EMBEDDING_MODEL = "intfloat/e5-large-v2"
+    EMBEDDING_MODEL = "intfloat/e5-large-v2"
     
-    # embeddings = HuggingFaceEmbeddings(
-    #     model_name=EMBEDDING_MODEL,
-    #     model_kwargs={'device': 'cpu'},
-    #     encode_kwargs={'normalize_embeddings': True}
-    # )
+    embeddings = HuggingFaceEmbeddings(
+        model_name=EMBEDDING_MODEL,
+        model_kwargs={'device': 'cpu'},
+        encode_kwargs={'normalize_embeddings': True}
+    )
 
-    embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001")
+    # embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001")
     CHUNK_SIZE = 700
     CHUNK_OVERLAP = 150
     text_splitter = RecursiveCharacterTextSplitter(chunk_size=CHUNK_SIZE,chunk_overlap = CHUNK_OVERLAP)