furst/llm-workshop
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Remove local_files_only=True from query_hybrid.py

Browse source 
The argument forces a strict deprecation path in newer
sentence-transformers/transformers that ignores cache_folder. Removing
it lets the script find the locally cached embedding model the same
way 03-rag's query.py does.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This commit is contained in:
Eric Furst 2026-05-01 12:13:34 -04:00
commit a1f9d4d5ed
1 changed files with 3 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

5
04-semantic-search/query_hybrid.py
View file

@ -20,7 +20,8 @@
# at import time.
import os
os.environ["TOKENIZERS_PARALLELISM"] = "false"
os.environ["SENTENCE_TRANSFORMERS_HOME"] = "./models"
os.environ["SENTENCE_TRANSFORMERS_HOME"] = os.path.abspath("./models")
os.environ["HF_HUB_CACHE"] = os.path.abspath("./models")
os.environ["HF_HUB_OFFLINE"] = "1"
from llama_index.core import (
@ -41,7 +42,7 @@ import sys
#
# Embedding model (must match build_store.py)
EMBED_MODEL = HuggingFaceEmbedding(cache_folder="./models", model_name="BAAI/bge-large-en-v1.5", local_files_only=True)
EMBED_MODEL = HuggingFaceEmbedding(cache_folder="./models", model_name="BAAI/bge-large-en-v1.5")
# LLM model for generation
LLM_MODEL = "command-r7b"