Products

PuzzleLib neural network framework is designed for building, training, and operating neural networks. The framework supports all cutting-edge types of neural networks and utilizes GPU accelerators.

We have developed a solution for classifying natural language text with the help of neural networks.

We develop algorithms for detecting and classifying objects in photos and videos. Our detectors perform at an average speed of 100 frames per second for a 640х480 video stream. Detection accuracy and exhaustiveness depend on the specific task and the required detection speed.

Our deep learning framework can operate on the Jetson TX2 or Xavier embedded computers. This allows us to launch neural network solutions in robots, cars, and drones.

Our PuzzleLib deep learning framework supports neural network integration for Apple devices (starting from iPhone 5s and up).

You can use PuzzleLib Mobile for building apps with augmented reality and apps for data recognition in photos and videos.