given a tensorflow model how do i convert it into pytorch
You could rewrite the model by porting each layer and verifying its functionality e.g. by making sure the random initialization of each parameter and buffer is equal.
The layer names should be quite similar for the majority of layers. I don’t know if there are any useful tools to automate this process.