This repository contains code for LARS (Layer-wise Adaptive Rate Scaling) based on Large Batch Training of Convolutional Networks implemented in PyTorch.
- Python 3.6
- PyTorch 1.0
from lars import LARS
optimizer = torch.optim.LARS(model.parameters(), lr=0.1, momentum=0.9)
optimizer.zero_grad()
loss_fn(model(input), target).backward()
optimizer.step()
