mindspore/mindspore/lite

查看中文

What Is MindSpore Lite

MindSpore lite is a high-performance, lightweight open source reasoning framework that can be used to meet the needs of AI applications on mobile devices. MindSpore Lite focuses on how to deploy AI technology more effectively on devices. It has been integrated into HMS (Huawei Mobile Services) to provide inferences for applications such as image classification, object detection and OCR. MindSpore Lite will promote the development and enrichment of the AI software/hardware application ecosystem.

For more details please check out our MindSpore Lite Architecture Guide.

MindSpore Lite features

1. Cooperative work with MindSpore training

   • Provides training, optimization, and deployment.
   • The unified IR realizes the device-cloud AI application integration.

2. Lightweight

   • Provides model compress, which could help to improve performance as well.
   • Provides the ultra-lightweight reasoning solution MindSpore Micro to meet the deployment requirements in extreme environments such as smart watches and headphones.

3. High-performance

   • The built-in high-performance kernel computing library NNACL supports multiple convolution optimization algorithms such as Slide window, im2col+gemm, winograde, etc.
   • Assembly code to improve performance of kernel operators. Supports CPU, GPU, and NPU.

4. Versatility

   • Supports IOS, Android.
   • Supports Lite OS.
   • Supports mobile device, smart screen, pad, and IOT devices.
   • Supports third party models such as TFLite, CAFFE and ONNX.

MindSpore Lite AI deployment procedure

1. Model selection and personalized training

   Select a new model or use an existing model for incremental training using labeled data. When designing a model for mobile device, it is necessary to consider the model size, accuracy and calculation amount.

   The MindSpore team provides a series of pre-training models used for image classification, object detection. You can use these pre-trained models in your application.

   The pre-trained models provided by MindSpore include: Image Classification and Object Detection. More models will be provided in the feature.

   MindSpore allows you to retrain pre-trained models to perform other tasks. For example: using a pre-trained image classification model, it can be retrained to recognize new image types. See Retraining.

2. Model converter and optimization

   If you use MindSpore or a third-party model, you need to use MindSpore Lite Model Converter Tool to convert the model into MindSpore Lite model. The MindSpore Lite model converter tool provides the converter of TensorFlow Lite, Caffe, ONNX to MindSpore Lite model, fusion and quantization could be introduced during convert procedure.

   MindSpore also provides a tool to convert models running on IoT devices .

3. Model deployment

   This stage mainly realizes model deployment, including model management, deployment, operation and maintenance monitoring, etc.

4. Inference

   Load the model and perform inference. Inference is the process of running input data through the model to get output.

   MindSpore provides a series of pre-trained models that can be deployed on mobile device example.