====== AI Training ====== This page describes how to train personal AI-detectors from reality data and related annotations {{orbit_desktop:how_to_find.png?25&nolink |}} Main Toolbar > Extract > AI Training ===== Concepts ===== ==== iTwin ==== An iTwin is a virtual replica of a physical assets powered by Bentley Systems. It will host all reality data linked to infrastructure project and will allow the start of AI-training job. ==== Training job ==== A training job will create an AI-Detector to be used in AI Detection sidebar. Relying on reality data (point cloud or images) and related annotations, it will be executed on Bentley Cloud Service. ==== Detectors ==== A detector is specific to a certain job type. The quality of the detection will depend on the similarity between your dataset and the training dataset’s description. A detector is a frozen model. Running on your reality data through an iTwin Capture Analysis job, it will automatically recognize elements of interest. Personal detectors can be trained from AI Training sidebar and a list of generic/pre-trained detectors is available here: \\ https://communities.bentley.com/products/itwin-capture/w/itwin-capture-wiki/66882/itwin-capture-detectors-download-page ===== Sidebar ===== {{ :254:desktop_ext:itc_manage_and_extract_244_aitraining_sb.png?direct&200|}} ==== iTwin ==== AI Training occurs on Bentley Cloud Service. An iTwin association is required to ensure reality data and related annotations are accessible to AI Training engine. ==== Upload ==== After iTwin association, select relevant reality data. If not available, open dedicated procedure to upload reality data Select it in AI Training sidebar ==== Cloud Reality Data ==== Select Reality Data you want your training job to un on. ==== Review training ==== Review training to ensure reality data is consistent for a training job ==== Initiate training ==== Trigger creation of specific cloud files that will enable training-job launch ==== Inform Bentley Support ==== Inform Bentley support Do not modify auto-generated email and send as is. Reality data and related annotations will be picked up by AI Training engine. An AI-detector will be shared by email at the end of the procedure. {{ :254:desktop_ext:itc_manage_and_extract_244_aitraining_email.png?direct&800 |}}