Deploy and Inference
The Deploy and Inference step allows users to run their build and perform inference on the selected remote device.
Note that this step is only unlocked when Step 3: Upload Input Data, Step 4: Pre-Processing Script (only if BYOM), and Step 5: Post-Processing Script (only if BYOM) are completed.
Starting the inference process
-
Review your customized hardware-specific settings in the configuration summary pane. A sample is shown below, but these parameters should match the selection from the Configure & Compile step.
-
(Optional) If you want to make changes, click Edit Parameters. This action will redirect the user to Step 2 – Configure and Compile.
-
Review all of your uploaded files that will be included in this particular inference session.
-
(Optional) If you want to make any edits, click Edit Files. This action will redirect you to Step 3 - Upload Input Data.
-
After you have reviewed the configuration settings and the input files, click Start Inference.
Note
Your request will be added to the queue. Queue times may vary depending on hardware availability. The inference status will change to Running once the request is processed by the remote board farm.
Take note of the Inference Statuses:
-
Failed: Errors must be resolved before re-sending the request. All errors are shown in the logs pane.
-
Aborted: Your request will be removed from the queue and/or any ongoing progress on the remote device will be lost. The user must click Start Inference to make a new request.
-
Completed: A new entry is added to the Inference History.
-
-
Once inference is complete, click Next to view the results.
Note
The Next button remains disabled until the inference process is successfully completed.
Logs pane
Detailed logs are displayed in the bottom pane, providing a summary of background activities and any errors encountered. These logs can be collapsed or expanded and are useful for troubleshooting and understanding what went wrong during compilation.




