Right now we have a algorithm base on pyTorch , we run this model in the I.MX8M mini and I.MX8M plus , they are also slow. When can we get the Deepview SDK ?
As I know , with using the Deepview SDK , we can convert the poTorch model to the converted model which can be accelerate computation by I.MX8M puls NPU .
As I know, Deepview SDK will release at July this year.
If we can share the algorithm source code to you , can you help to convert it for test firstly ?
Good!