Skip to main content

Conda Environment Setup

Click "Setup Env", enter a name, choose a Python version (3.9 or 3.10 recommended), and choose the environment type.

IMPORTANT

A Conda environment is mandatory before running preprocessing, training, or generating deployment packages.

Audio Classification​

  • Click "Browse Folder" and select the folder named exactly: syntiant-ndp-model-converter
  • Click "Setup Env" → select Audio/Gesture (Syntiant) → create environment
  • An external terminal opens to run the Syntiant converter install script
  • After install completes, return to the Dashboard and click "Installation Completed"
NOTE

Each version of the Syntiant Model Converter requires a separate environment setup.

Gesture Recognition​

Same steps as Audio Classification — the Syntiant converter folder is also required for Gesture Recognition projects.

Image Classification​

  • No Syntiant converter folder required
  • Click "Setup Env" → select Image (YOLO) → enter name → choose Python 3.10 → click Create
  • An external terminal installs: ultralytics, torch, torchvision, onnx, onnxruntime, opencv-python, numpy, matplotlib, scikit-learn
  • Return to the Dashboard and click "Installation Completed" → "Environment is ready!" confirmation appears
NOTE

If the Console Log shows "⚠ Conda not found on system", install Miniconda first before attempting environment setup.