# Frequently Asked Questions

# Q: Why doesn't the generated Edge code (.c/.h) build in my compiler?

A: Configure your compiler for the C99 and ANSI C standard. Make sure that you don't treat warnings as errors.

# Q: Why is nothing added to the batch import when I select my sessions as "Additional tracks..."?

A: If you edit any of your .imsession files you should re-import them into your database before sending a job to the Training Service. If you don't there will be a mismatch between your session ID's in the database and the session ID's of your actual sessions, making it impossible for the Batch Import Tool to link your downloaded predictions to your local sessions.

# Q: How do I activate experimental features?

A: Ctrl+Shift+D or File -> Preferences -> Plugins -> Enable Experimental Features. Some features require that you restart the Studio for the change to have effect. Note: These features are provided without support.

# Q: My jobs seems to run at different speeds?

A: The AI Training Service is free to assign jobs to different compute resources. This might have this effect.

# Q: I cannot delete my "Edge" directory because a file is being used by another process.

A: A common cause of this error is that the Edge directory is your current working directory in Cygwin. Close cygwin and try again.

# Q: Does the time_out variable in the C code give you the first or last timestamp in the window for the prediction?

A: The time_out variable is an array containing all timestamps that was input to the model to form the current prediction. In most cases, the first timestamp in the array is the first timestamp of the input for the prediction and the last array entry is the last timestamp of the input for the prediction.

# Q: Can I deploy the Python/.h5 models?

A: Yes, it is possible to deploy the python models. This is especially useful for performing live testing before putting the AI application on the edge device and we have a special tool for this purpose. Please contact us for more details.