Nota Inc. provides on-device AI solutions that remove the need for any server or cloud computing. With our deep learning model compression techniques, including quantization, pruning, knowledge distillation, and filter decomposition, vision-based models can be stored individually on small edge devices without performance degradation. Our compression removes latency, server/network cost, and privacy problems, while still achieving state-of-the-art performance at object detection, identification, tracking, etc.
Nota recently succeeded in developing the Automatic Model Compression (AMC) platform that automatically compresses the size of computer vision models based on the optimization of the model, target hardware, and the dataset. AMC will eventually free human labor while achieving a higher compression ratio, greater accuracy, and faster inference. We believe our compression platform will accelerate the deployment of on-device computer vision solutions to every aspect of society.