I have explored edge computation for AI and I came across multiple library or framework which can help to convert model into lite format which is suitable for edge devices.
Tensorflow Lite
will help us to convert Tensorflow
Model into Tensorflow lite.
Or Intel OpenVino
will optimise the model for edge devices.
Questions
- If we have library to optimise the model for edge devices, after conversion will it cause any accuracy issue. performance will get increase but accuracy of that model will get decrease.
- If we have such technology to optimise model performance and wont get accuracy decrease then why people wont prefer
tensorflow Lite model
overtensorflow model
.