Hi, thanks for this great project. I've been learning your code and have a question here.
In the original bert code, we need to calculate the input_mask feature when creating the inputs.
But in your version, the [input_mask] seems uncalculated.
So I'm very confused about how it works.
Looking forward to your reply.
Hi, thanks for this great project. I've been learning your code and have a question here. In the original bert code, we need to calculate the input_mask feature when creating the inputs. But in your version, the [input_mask] seems uncalculated. So I'm very confused about how it works. Looking forward to your reply.