Model Test

run flask

flask --app createEmbeddings2 run

--host=0.0.0.0

curl http://127.0.0.1:5000/embedding -H "Content-Type: application/json" --request POST --data '{"text":"xyz"}'

local model store

find ~/.cache/huggingface/hub/

python3 -c 'from sentence_transformers import SentenceTransformer; SentenceTransformer("all-MiniLM-L6-v2", cache_folder="./app/artefacts")'

DB

• solr support no multivalued fields in 9.7
• chromadb may be limited in scaling
• pinecone is online
• milvus might be the solution for billions of embeddings