Open water-in-stone opened 22 hours ago
Hi @water-in-stone , It looks like you are not persisting your vector store. It means that the vector store configuration that you are using currently:
"vector_store": {
"provider": "qdrant",
"config": {
"collection_name": "test",
"embedding_model_dims": 3072,
},
}
will create a temporary collection in tmp/qdrant
that you will loose as soon as you close your program agent.py
. I recommend you to use docker container configuration by using these commands:
docker pull qdrant/qdrant
docker run -p 6333:6333 -p 6334:6334 \
-v $(pwd)/qdrant_storage:/qdrant/storage:z \
qdrant/qdrant
and then use these configurations:
"vector_store": {
"provider": "qdrant",
"config": {
"host": "localhost",
"port": 6333,
}
},
🐛 Describe the bug
My source code is shown above and stored in a file named 'agent.py'. I can input a question and get the correct answer after I execute python agent.py. However, I cannot get the correct answer when I execute agent.py again with the same question. Is Mem0 only able to store data in memory and not on the hard disk?