Supported Task Types in QpiAI Pro
QpiAI Pro supports the following task types, each requiring a specific Domain and Sub-domain selection during workspace creation:
1. Computer Vision (CV)
CV tasks work with image and video data and include:
Object Detection (OD) — detect and locate objects in images using bounding boxes
Image Segmentation — identify and segment objects at pixel level
Video Tracking — detect and track objects across video frames over time using text-prompted targets, producing labeled bounding boxes with persistent object IDs throughout the video
Medical Image Segmentation — specialized segmentation for medical imaging (e.g., MRI, pathology slides). Select Healthcare as your Domain with Sub-domains such as Medical Imaging & Radiology or Pathology & Histology to unlock medical segmentation tasks.
2. LLM (Large Language Models)
LLM tasks work with text data and are used to build domain-specific AI assistants, Q&A bots, and language-based models. Includes:
Instruction Tuning — train using prompt and response pairs
Preference Tuning — train using prompt, chosen, and rejected response triplets
The task type available to you depends entirely on the Domain and Sub-domain you select when creating your workspace. CV tasks are unlocked by selecting a CV Domain and Sub-domain, while LLM tasks are unlocked by selecting a GenAI Domain and Sub-domain. Refer to Appendix A: Domain and Sub-domain Reference for the complete list.