Sign up to take part
Registered users can ask their own questions, contribute to discussions, and be part of the Community!
Registered users can ask their own questions, contribute to discussions, and be part of the Community!
We have a model which should predict in real time.
Since we have a lot of custom data pre-processing steps in the flow, we have included all the steps as scripts in the library and have imported the scripts using a python function.
We have deployed an API service with a python function end point.
When trying to get a prediction from a model which has been deployed to the flow, we are seeing this error.
Failed: Failed to run function : <class 'FileNotFoundError'> : [Errno 2] No such file or directory: '/data/dataiku/data_dir/saved_models/ML_EDA/<ModelID>/versions/16635986036/core_params.json'
Can we include the model in the Deployer node?
We are not seeing the model export option to possibly add the model into a managed folder.
Can you please suggest an alternative way to achieve this?
Thanks
--------------------------------------------------------------------------------
I was not able to post a reply to this thread.
We were able to resolve this issue by deploying the model as a prediction end point under the same parent service and invoking the prediction service from the python code of the other end point using the code snippet shown below:
from dataiku.apinode import utils
apinodeclient = utils.get_self_client()
prediction = apinodeclient.predict_record("<EndPoint>","<features>", with_explanations=True)
Operating system used: Linux