Hi,
Recently we started with i.mx8m plus SoC for ML tasks and I"m quite curious about some roadmap for TensorFlow-lite versions support. About a week or two there was released version 2.2.0, which replaced 2.1.0. However, all our models are developed/trained in 2.3.0.
Could you please shed some light on when tf-lite 2.3.0 will be integrated into BSP?
Is there any chance to do it on my own? (this is not about how to modify Yocto recipes, but rather if there is a special imx version of tf-lite I can integrate into BSP)