MindSpore: An Open-Source Deep Learning Training Framework For Mobile, Edge And Cloud Scenarios

This lightweight framework is ready to give competition to Google's TensorFlow, and Facebook's PyTorch, and it can scale across devices, cloud, and edge environments. One of the key competitive advantages with'Mindspore' is that it uses 20% fewer codes that its competitors for a function like NLP (Natural language processing). Apart from codes, it can also support parallel training to save training time across hardware. Huawei developed this framework with support from partners like the University of Edinburgh, Peking University, Imperial College London, and robotics startup Milvus. Mindspore maintains and preserves sensitive data by not processing any data itself but ingests only the gradient and model information that has already been processed.