Example for loading llama 7b or bert model into tvm api

Questions
1

import tvm from tvm.driver import tvmc

model = tvmc.load(‘llama-7b.pth’)

package = tvmc.compile(model, target=“llvm”)

trying to load llama-7b model but getting input shape error, shouldn’t input shape be variable token for LLms? any example how bert model is loaded and optimized on tvm?