Recommendation about Module Predict Control for time series dataset

I have a time series iot dataset and I am interested in NVIDIA for building a module predict control for prediction. However, I am unable to find anything like it. Any recommendations!

You can explore NVIDIA RAPIDS, NVIDIA DeepStream, and the NVIDIA Triton Inference Server.
Please refer to the above product documentation for more details.

1 Like

Thank you for the recommendation. I am thinking of using a jetson device for this use i.e. developing a module predicitive control for a PLC based control system. I have some devices shortlisted and need some recommendation:
1- NVIDIA Jetson Nano Development kit
2- AGX Xavier Industrial
3- AGX Xavier

I want to accumulate TB worth of data generate by the sensors, build model predictive control and perform inference that can communicated to control system. Should I use one of the above device to solely build end-to-end application (accumulating data, training and inference)? Any other possible solution by NVIDIA relevant to my use case (not a computer vision problem).


We are moving this post to the Jetson related forum to get better help.

Thank you.

You may refer to below links to make your decision:
Jetson Benchmarks | NVIDIA Developer
Jetson FAQ | NVIDIA Developer

@sale2 I have some basic time-series forecasting/prediction examples using PyTorch here, but it is introductory and far from SOTA so you will probably want to look for research or other projects in the area:

Also since I made that, a new package called pytorch-forecasting as come out: PyTorch Forecasting Documentation — pytorch-forecasting documentation

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.