Skip to content

Getting Started

Getting Started

The Python SDK enables you to integrate AI Video & Image Analysis into your applications or use it directly.

Prerequisites

Activate your license on the target machine where the SDK is installed.
See: License Activation for instructions.

Server Requirements

To ensure optimal performance, your server should meet the following specifications. Flash Attention support is required for maximum efficiency.

Operating System

  • Linux (Ubuntu 22.04+ recommended)

CUDA

  • CUDA >= 12.3.
  • GPU Architecture: Ampere, Blackwell
  • GPU Memory: 20 GB++
  • Storage for SDK and dependency: 25GB

Software Dependencies

  • Python 3.10 or higher
  • NVIDIA drivers compatible with your CUDA version

Note: GPU Memory Usage

GPU and system memory usage depend on the configured input size. Larger inputs consume more of both, and may cause out‑of‑memory errors if hardware capacity is insufficient. It is important to consider the server’s hardware limits and the requirements of the specific scenario when selecting input settings. The hardware specifications listed above provide a baseline for processing small batch sizes of video frames; higher-capacity hardware may be necessary for larger batches.