[AutoTVM] Adjustment for new OP

Hello everyone. Im still rather new to TVM, so my understanding of how TVM works under the hood is quite restricted. So my goal is to add the resampler (https://www.tensorflow.org/addons/api_docs/python/tfa/image/resampler) OP to TVM from Tensorflow frontend. As a guideline i used the implementation of Upsampling, since their logic and computation is quite similar. I was able to get my resampler OP running by (to my understanding) implementing the Frontendside, the Relay Call Node association and the Topi logic. Running a dummy model actually worked pretty good, i was able to get correct results through TVM. Next of trying to use AutoTVM to optimize the schedule of my OP does not work (it does also not work when feeding a dummy Upsampling Tensorflow model). Now to my understanding i need to specify configs for the schedule (or OP overall) so that TVM recognizes it and can optimize it. Sadly i am not getting my head around different examples, e.g. starting python/tvm/relay/op/nn/_nn.py where we define the OP type and schedule). So my question is: which steps do i need to follow (e.g. which things do i need to define/implement in which locations of the TVM stack) in order to allow my own OP to be optimized through AutoTVM?

I would be very thankful for some suggestions. Best regards, Knight3