A Python library for storing and managing machine learning models in MongoDB using GridFS.
- Store machine learning models in MongoDB's GridFS.
- Easily connect to MongoDB using
MongoDBConnector.
To install the library locally, follow these steps:
-
Create your python environment and activate it.
-
Clone the repository:
git clone https://github.com/tetaud-sebastien/model-registry cd model-registry -
Install the library:
pip install -e .
from model_registry.connector import MongoDBConnector
username = "username"
password = "password"
host = "ip_address"
db_name = "db_name"
# Connect to MongoDB
connector = MongoDBConnector(username, password, host, db_name)
client = connector.connect()from model_registry.store import store
collection = "your_collection"
model_path = "path/to/model.onnx"
# Metadata for the model
metadata = {
"project_name": "ssh_mapping",
"model_architecture": "AutoencoderCNN3D",
"model_version": "0",
}
# Connect to MongoDB
connector = MongoDBConnector(username, password, host, db_name)
client = connector.connect()
# Store the model in MongoDB GridFS
success = store(client=client, db=db_name, collection=collection, metadata=metadata, model_path=model_path)Feel free to open an issues whether you want for reporting issues, improving documentation, or contributing code, your input is valuable.