Runtime Environment
Managing Python runtime installation and GPU acceleration setup
The Runtime Environment screen manages the self-contained Python installation that Digit Desktop uses for all image processing, AI inference, and data export operations. This is typically the first screen you visit after installing the application.
Prerequisites
An internet connection for downloading the runtime.
Administrator privileges are not required — the runtime installs to the user data folder.
Screen Layout
The screen displays:
Runtime Status: Current installation state and version.
Action Buttons: Install, Update, or Check GPU.
Progress Area: Live log output and progress bar during operations.
Runtime Status
Idle
No operation in progress
Ready for commands
Not Installed
Python runtime is missing
Click Install Runtime
Outdated
A newer version is available
Click Install Runtime to update
Installed
Runtime is present and up-to-date, version displayed
No action needed
Running
An operation is in progress
Wait for completion
Success
Last operation completed successfully
Proceed to use the application
Error
Last operation failed
Review the log output for details
Step-by-Step Walkthrough
1. Check Current Status
When you navigate to the Runtime Environment screen, the application automatically checks:
Whether the Python runtime is installed.
The installed version number.
Whether an update is available.
2. Install or Update the Runtime
Click Install Runtime (the button label changes to "Update Runtime" when an older version is detected).
The application downloads a portable Python distribution from Csmart's cloud storage (Google Cloud Storage).
A progress bar shows the download and extraction percentage.
The log panel streams real-time output including download speed, file extraction, and validation steps.
The sidebar is locked during installation to prevent navigation.
The runtime is approximately 1-2 GB in size. Download time depends on your internet speed. The installation is fully self-contained and does not modify your system Python or PATH.
Do not close the application during installation. If the process is interrupted, you can restart it by clicking Install Runtime again.
3. Verify GPU Availability
After the runtime is installed, click Check GPU to verify NVIDIA GPU access:
The application spawns a Python diagnostic script.
The script checks for CUDA availability and reports:
GPU name and model.
Available VRAM.
CUDA version.
Whether ONNX Runtime can use the GPU.
Results are displayed in the log panel.
GPU acceleration is optional but significantly speeds up AI inference. If no GPU is detected, the application falls back to CPU-based inference.
4. Verify Python Path
The screen displays the resolved Python executable path, confirming where the runtime is installed. In development mode, this points to the local virtual environment. In production, it points to the packaged runtime in the application's resources folder.
Error Recovery
If installation fails:
Review the error message displayed in the status area.
Check the log output for specific failure details.
Common issues include:
Network timeout during download — retry the installation.
Insufficient disk space — free up space and retry.
Antivirus blocking extraction — add an exception for the application's data folder.
Click Install Runtime again to retry.
Troubleshooting
Download fails or is very slow
Poor internet connection or firewall
Check your connection; add the application to firewall exceptions
"Not Installed" after successful install
Application restarted before extraction completed
Re-run the installation
GPU not detected
NVIDIA drivers missing or outdated
Install the latest NVIDIA drivers from nvidia.com
GPU detected but inference uses CPU
ONNX Runtime CUDA provider not available
Reinstall the runtime to get matching CUDA libraries
Status stuck on "Running"
Process stalled during installation
Close and reopen the application; retry installation
Insufficient disk space
Runtime requires 1-2 GB of free space
Free up disk space and retry
Last updated