Object detection was the first problem type supported when Nilvana™ vision studio was released last year. Since then, we have conducted extensive research and gathered client feedback from many individuals. Given the recent presentation of multiple new one-stage algorithms by various researchers, it is time to upgrade our current implementations. Support for multiple GPU training is one of the most commonly requested feature improvements, even though most users only need one graphics card. However, we will include this option in the upcoming release to satisfy customer demand.
Although we are a tiny team, we are passionate about exploring new technologies and hope that the tools we develop will benefit humanity. If you have any computer vision related questions, please feel free to contact us, our technical staff will try its best to come up with a feasible solution.