vector db retrain

app.py

@@ -22,7 +22,7 @@ def retrieve_documents():
         #api_key=HF_TOKEN, model_name="local:BAAI/bge-m3")
     db = Chroma(persist_directory="./db",
             embedding_function=embeddings)
-    retriever = db.as_retriever(search_kwargs = {"k":1})
+    retriever = db.as_retriever(search_kwargs = {"k":3})
     return retriever
 
 @st.cache_resource()
@@ -42,8 +42,8 @@ def create_chain(_retriever):
 
     llm = HuggingFaceEndpoint(
     endpoint_url = "https://api-inference.huggingface.co/models/mistralai/Mistral-7B-Instruct-v0.2",
-    max_new_tokens=512,
-    top_k=1,
+    max_new_tokens=2048,
+    top_k=3,
     top_p=0.9,
     temperature=0.5,
     repetition_penalty=1.1,

data/Sungwon_Kim_ML_DL_Intro_together.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2bf23de972563f545577fc3ba526a7095ade61ac4c9ff31a3f0b0a431fe0a9c6
+size 166124

db/{727833e8-2ded-474a-9aa7-68262d54e1bf → 7ff82419-a056-482d-b47b-bf20f0416667}/length.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c6288e3187b841602a59f0ecf19ab89046d64462debfe6e37465c2a39a436af5
+oid sha256:b72fff46a00133ed25e0295587e9d32f02f68c7dc9249872d219b49934c119de
 size 4000

db/chroma.sqlite3

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cc61685aeb6a95f279aac19f5b6fe590bb923ec59e754e48894f4196b92d87ed
-size 450560
+oid sha256:63e5791397a6bc2efe890213198b6bd62742f7d4342c90eed8ecec781a11824f
+size 270336

gen.py

@@ -7,7 +7,6 @@ from langchain_community.embeddings import HuggingFaceInferenceAPIEmbeddings
 from langchain_community.vectorstores import Chroma
 import streamlit as st
 
-
 HF_TOKEN = st.secrets["HF_TOKEN"]
 
 def persist_dir(file_path):