I have the following questions regarding CV or ComputerVision:
(1) I would like to know what are the possible ways in order for ComputerVision to work offline?
(2) Is it possible to stop using ComputerVision in the first place and use an alternative? Can you tell me possible alternatives?
(1) ComputerVision : you can explore on-device computer vision libraries and frameworks such as OpenCV or custom machine learning models deployed locally.
(2) Yes, it’s possible to use alternatives to ComputerVision, including OpenCV, TensorFlow, PyTorch, and custom machine learning models, depending on your specific use case and requirements.
(1) ComputerVision : you can explore on-device computer vision libraries and frameworks such as OpenCV or custom machine learning models deployed locally.
(2) Yes, it’s possible to use alternatives to ComputerVision, including OpenCV, TensorFlow, PyTorch, and custom machine learning models, depending on your specific use case and requirements.
Thank you for your response.
I have a follow-up question. Before these alternative learning models can be used by UiPath, I will need to deploy an on-premise Orchestrator? Or are there alternative ways?
Basically there are 3 available CV deployments, all free of charge:
Cloud CV, which uses our servers. To use this method, you’ll need to log into Automation Cloud and get the API key from Admin > Licenses > Services
On-prem CV server – requires a dedicated machine with GPU
Local/robot CV – detections are performed on the robot that runs the automation, so no server required, but it slightly slower, has a lower accuracy, and a few features are unavailable.
With on-premise version of the ML model which allows AI Computer Vision to be used without internet connection. If you’re interested, please email insider-preview@uipath.com and we’ll follow up soon with the details/documentation
For many of the alternative learning models and AI services that can be integrated with UiPath, deploying an on-premise Orchestrator may not be a strict requirement. The need for an Orchestrator largely depends on the specific AI service or library you intend to use and the use case you have in mind.