Huawei introduces MindSpore an AI computing framework last year. During the launch event, the company has stated that an AI computing framework MindSpore will be open source in the first quarter of 2020.
Huawei today at the HDC Cloud 2020 officially announced that the Huawei MindSpore is now open source. Huawei MindSpore is an AI computing framework that provides unified APIs and end-to-end AI capabilities for model development, execution, and deployment in all scenarios.
Using a distributed architecture, MindSpore leverages a native differentiable programming paradigm and new AI native execution modes to achieve better resource efficiency, security, and trustworthiness. Meanwhile, MindSpore makes full use of the computing power of Ascend AI processors and lowers the entry requirements of industry AI development, bringing inclusive AI faster to reality.
Huawei MindSpore Features:
- Development Friendliness
- Execution Efficiency
- Deployment Flexibility
Development Friendliness:
- Automatic differential: unified programming of network and operators, native expression of functions and algorithms, and automatic generation of inverse network operators.
- Automatic parallelization: achieving the best model parallelization with automatic model partitioning.
- Automatic optimization: using the same code for dynamic and static computation graphs.
Execution Efficiency:
- On-device execution, making full use of the computing power of Ascend AI processors.
- Pipeline optimization, maximizing parallel processing linearity.
- Deep graph optimization, automatically adapting to the AI core computing power and precision.
Deployment Flexibility:
- On-demand collaborative computing across the cloud-edge-device, better-protecting privacy.
- A unified architecture for the cloud-edge-device, implementing one-time development and on-demand deployment.
Huawei HDC.Cloud 2020 Latest News: