Optimizers
This contains a set of optimizers.
wrap_optimizer
wrap_optimizer (opt, **kwargs)
You can natively use any of the optimizers included in the fastai library. You just need to pass it to the learner as the opt_func.
In addition, you will be able to use any of the optimizers from:
- Pytorch
- torch_optimizer (https://github.com/jettify/pytorch-optimizer). In this case, you will need to install
torch-optimizer
first)
Examples of use:
adamw = wrap_optimizer(torch.optim.AdamW)
import torch_optimizer as optim
adabelief = wrap_optimizer(optim.AdaBelief)
If you want to use any these last 2, you can use the wrap_optimizer function. Here are a few examples: