Generate a metadata.json
Generate a metadata.json
file to provide information about the model and training data, in order to enable features related to monitoring of prediction log data (e.g. drift monitoring). More information about the metadata.json
can be found in the Repository requirements
An example of how to generate a metadata.json
for your model:
from sklearn import datasets
from deeploy.common.functions import generate_metadata
from deeploy.client import Client
from deeploy.enums.metadata import ProblemType
data = datasets.load_iris()
df_train = pd.DataFrame(data=data.data, columns=data.feature_names)
metadata = generate_metadata(df_train, problem_type=ProblemType.CLASSIFICATION)
# Save to a 'metadata.json' file
client = Client(
host="app.deeploy.ml",
workspace_id="b6d8c781-2526-4e03-9b43-example",
access_key="DPAexample",
secret_key="example",
)
client.generate_metadata_json("./directory/", metadata)