Chinese robotics firm AgiBot has unveiled its AI-powered humanoid robots training dataset.
According to SiliconANGLE on Monday, the dataset encompasses a variety of training sets crucial for robots to execute human-like activities.
AgiBot, a rising competitor to industry giants like Tesla and Boston Dynamics, is developing humanoid robots capable of performing basic human tasks such as folding clothes, organizing items, cleaning, and doing laundry. To accelerate this ambitious project, the company is set to launch “AgiBot World,” an extensive open-source dataset offering more than 1 million different training sets for 100 robots.
The dataset includes tasks in various environments, including homes, restaurants, industrial settings, offices, and supermarkets, and is used to train robots to mimic human movements. Moreover, the robots trained with this dataset can collaborate with human workers and other robots.
AgiBot World will be available on Hugging Face and GitHub for robotics AI developers and academic researchers to utilize and customize their own AI models for humanoid robots.
Most Commented