What is the best approach to convert a pytorch model to TVM?

Questions
#1

I have successfully trained a pytorch model and i am facing issues of slow inferencing, What is the bes approach for faster inferencing? and how do i convert a pytorch trained model to TVM ? I am new with TVM.

#2

https://tvm.apache.org/docs/tutorials/frontend/from_pytorch.html#sphx-glr-tutorials-frontend-from-pytorch-py

https://tvm.apache.org/docs/tutorials/frontend/deploy_object_detection_pytorch.html#sphx-glr-tutorials-frontend-deploy-object-detection-pytorch-py

#3

Can i use torch==1.6.0 and torchvision== 0.7 ? becasue i trained my model using these versions…

#4

I remember seeing this PR, which is not merged yet, mentioning torch 1.6. Maybe @masahi can comment here.