DL-Art-School/codes/trainer/optimizers/larc.py
James Betker 036684893e Add LARS optimizer & support for BYOL idiosyncrasies
- Added LARS and SGD optimizer variants that support turning off certain
  features for BN and bias layers
- Added a variant of pytorch's resnet model that supports gradient checkpointing.
- Modify the trainer infrastructure to support above
- Fix bug with BYOL (should have been nonfunctional)
2020-12-23 20:33:43 -07:00

110 lines
4.1 KiB
Python

import torch
from torch import nn
from torch.nn.parameter import Parameter
class LARC(object):
"""
:class:`LARC` is a pytorch implementation of both the scaling and clipping variants of LARC,
in which the ratio between gradient and parameter magnitudes is used to calculate an adaptive
local learning rate for each individual parameter. The algorithm is designed to improve
convergence of large batch training.
See https://arxiv.org/abs/1708.03888 for calculation of the local learning rate.
In practice it modifies the gradients of parameters as a proxy for modifying the learning rate
of the parameters. This design allows it to be used as a wrapper around any torch.optim Optimizer.
```
model = ...
optim = torch.optim.Adam(model.parameters(), lr=...)
optim = LARC(optim)
```
It can even be used in conjunction with apex.fp16_utils.FP16_optimizer.
```
model = ...
optim = torch.optim.Adam(model.parameters(), lr=...)
optim = LARC(optim)
optim = apex.fp16_utils.FP16_Optimizer(optim)
```
Args:
optimizer: Pytorch optimizer to wrap and modify learning rate for.
trust_coefficient: Trust coefficient for calculating the lr. See https://arxiv.org/abs/1708.03888
clip: Decides between clipping or scaling mode of LARC. If `clip=True` the learning rate is set to `min(optimizer_lr, local_lr)` for each parameter. If `clip=False` the learning rate is set to `local_lr*optimizer_lr`.
eps: epsilon kludge to help with numerical stability while calculating adaptive_lr
"""
def __init__(self, optimizer, trust_coefficient=0.02, clip=True, eps=1e-8):
self.optim = optimizer
self.trust_coefficient = trust_coefficient
self.eps = eps
self.clip = clip
def __getstate__(self):
return self.optim.__getstate__()
def __setstate__(self, state):
self.optim.__setstate__(state)
@property
def state(self):
return self.optim.state
def __repr__(self):
return self.optim.__repr__()
@property
def param_groups(self):
return self.optim.param_groups
@param_groups.setter
def param_groups(self, value):
self.optim.param_groups = value
def state_dict(self):
return self.optim.state_dict()
def load_state_dict(self, state_dict):
self.optim.load_state_dict(state_dict)
def zero_grad(self):
self.optim.zero_grad()
def add_param_group(self, param_group):
self.optim.add_param_group(param_group)
def step(self):
with torch.no_grad():
weight_decays = []
for group in self.optim.param_groups:
# absorb weight decay control from optimizer
weight_decay = group['weight_decay'] if 'weight_decay' in group else 0
weight_decays.append(weight_decay)
group['weight_decay'] = 0
for p in group['params']:
is_bn_or_bias = (hasattr(p, 'is_bn') and p.is_bn) or (hasattr(p, 'is_bias') and p.is_bias)
if p.grad is None or is_bn_or_bias:
continue
param_norm = torch.norm(p.data)
grad_norm = torch.norm(p.grad.data)
if param_norm != 0 and grad_norm != 0:
# calculate adaptive lr + weight decay
adaptive_lr = self.trust_coefficient * (param_norm) / (
grad_norm + param_norm * weight_decay + self.eps)
# clip learning rate for LARC
if self.clip:
# calculation of adaptive_lr so that when multiplied by lr it equals `min(adaptive_lr, lr)`
adaptive_lr = min(adaptive_lr / group['lr'], 1)
p.grad.data += weight_decay * p.data
p.grad.data *= adaptive_lr
self.optim.step()
# return weight decay control to optimizer
for i, group in enumerate(self.optim.param_groups):
group['weight_decay'] = weight_decays[i]