This is an explanation on how to implement bert on different devices like CPU and GPU.
To check and install for cuda compatible devices, I think its [documentation](https://docs.nvidia.com/cuda/cuda-installation-guide-microsoft-windows/index.html) is far more appropriate.
Many other places where Bert models can be deployed majorly depends on the type of data that we choose. Like if we want to deploy it in an app or just as a part of biger model architecture for recomendation system or speech generation models. So this basic immplementation covers how we can fine tunning bert at its base level by freezing most of parameters and taking full benefit of transfer learning.