I modified the official bert code and just modified tupe stiamtor to normal estimator, I did not change the model architecture. When I loaded the pb file and run build, it failed. The error is below.
Incompatible broadcast type TensorType([?, 384, 2], float32) and TensorType([4, 2], float32)
The type inference pass was unable to infer a type for this expression.
This usually occurs when an operator call is under constrained in some way, check other reported errors for hints of what may of happened.
The type inference pass was unable to infer a type for this expression.
This usually occurs when an operator call is under constrained in some way, check other reported errors for hints of what may of happened.
The type inference pass was unable to infer a type for this expression.
This usually occurs when an operator call is under constrained in some way, check other reported errors for hints of what may of happened.
The type inference pass was unable to infer a type for this expression.
This usually occurs when an operator call is under constrained in some way, check other reported errors for hints of what may of happened.
note: run with TVM_BACKTRACE=1 environment variable to display a backtrace.
How can I fix it.