Need Guidance Regarding Model Deployment
Hi Team & Members,
Model Type(Auto Model) - Binary Logistic Regression (Labels to predict Approved/Denied)
(New to Rapid Minner)
Problem Statement:
My problem statement is that i want to deploy the model on a robot(Small Autonomous Vehicle) which will have no internet connection and only has CUDA C++ and Python installed on a windows OS. (Note Cannot Install Rapid Miner Software as hardware reservations for other processes are delicately calibrated).
Requirement:
I need to run the model through python (json/csv as input, passing other formats is also acceptable) and generate an output file.
Please help me with a way that i can achieve such a goal, any help is much appreciated.
(Also please help me understand how and where or if i should use pmml extension)
(Please check my attempt to export the auto model - production model processes to pmml)
Regards,
Yash Sharma
Model Type(Auto Model) - Binary Logistic Regression (Labels to predict Approved/Denied)
(New to Rapid Minner)
Problem Statement:
My problem statement is that i want to deploy the model on a robot(Small Autonomous Vehicle) which will have no internet connection and only has CUDA C++ and Python installed on a windows OS. (Note Cannot Install Rapid Miner Software as hardware reservations for other processes are delicately calibrated).
Requirement:
I need to run the model through python (json/csv as input, passing other formats is also acceptable) and generate an output file.
Please help me with a way that i can achieve such a goal, any help is much appreciated.
(Also please help me understand how and where or if i should use pmml extension)
(Please check my attempt to export the auto model - production model processes to pmml)
Regards,
Yash Sharma
Find more posts tagged with
Sort by:
1 - 1 of
11
Unfortunately, there is no easy way to do this cross-platform. RapidMiner is designed to be able to do the scoring deployment and to provide it as a web service. So if your device is not going to have access to the internet, probably your best solution will be to build your model directly in python so it can be deployed on your platform. Alternatively you could look at using a different modeling method such as a decision tree, which is more easily ported in terms of the code into other languages.