Hugging Face's logo

Bitsandbytes documentation

LARS

You are viewing main version, which requires installation from source. If you'd like regular pip install, checkout the latest stable version (v0.44.1).
Hugging Face's logo
Join the Hugging Face community

and get access to the augmented documentation experience

to get started

LARS

LARS (Layer-wise Adaptive Rate Scaling) is an optimizer designed for training with large batch sizes to accelerate training. LARS uses a separate learning rate for each layer instead of each parameter. The learning rate is calculated from a trust ratio between the weight and gradient norm in a layer. This helps calibrate a stable update size.

LARS

class bitsandbytes.optim.LARS

< >

( params lr momentum = 0 dampening = 0 weight_decay = 0 nesterov = False optim_bits = 32 args = None min_8bit_size = 4096 percentile_clipping = 100 max_unorm = 0.02 )

__init__

< >

( params lr momentum = 0 dampening = 0 weight_decay = 0 nesterov = False optim_bits = 32 args = None min_8bit_size = 4096 percentile_clipping = 100 max_unorm = 0.02 )

Parameters

  • params (torch.tensor) — The input parameters to optimize.
  • lr (float) — The learning rate.
  • momentum (float, defaults to 0) — The momentum value speeds up the optimizer by taking bigger steps.
  • dampening (float, defaults to 0) — The dampening value reduces the momentum of the optimizer.
  • weight_decay (float, defaults to 1e-2) — The weight decay value for the optimizer.
  • nesterov (bool, defaults to False) — Whether to use Nesterov momentum.
  • optim_bits (int, defaults to 32) — The number of bits of the optimizer state.
  • args (object, defaults to None) — An object with additional arguments.
  • min_8bit_size (int, defaults to 4096) — The minimum number of elements of the parameter tensors for 8-bit optimization.
  • percentile_clipping (int, defaults to 100) — Adapts clipping threshold automatically by tracking the last 100 gradient norms and clipping the gradient at a certain percentile to improve stability.
  • max_unorm (float, defaults to 0.02) — The maximum gradient norm.

Base LARS optimizer.

LARS8bit

class bitsandbytes.optim.LARS8bit

< >

( params lr momentum = 0 dampening = 0 weight_decay = 0 nesterov = False args = None min_8bit_size = 4096 percentile_clipping = 100 max_unorm = 0.02 )

__init__

< >

( params lr momentum = 0 dampening = 0 weight_decay = 0 nesterov = False args = None min_8bit_size = 4096 percentile_clipping = 100 max_unorm = 0.02 )

Parameters

  • params (torch.tensor) — The input parameters to optimize.
  • lr (float) — The learning rate.
  • momentum (float, defaults to 0) — The momentum value speeds up the optimizer by taking bigger steps.
  • dampening (float, defaults to 0) — The dampening value reduces the momentum of the optimizer.
  • weight_decay (float, defaults to 1e-2) — The weight decay value for the optimizer.
  • nesterov (bool, defaults to False) — Whether to use Nesterov momentum.
  • args (object, defaults to None) — An object with additional arguments.
  • min_8bit_size (int, defaults to 4096) — The minimum number of elements of the parameter tensors for 8-bit optimization.
  • percentile_clipping (int, defaults to 100) — Adapts clipping threshold automatically by tracking the last 100 gradient norms and clipping the gradient at a certain percentile to improve stability.
  • max_unorm (float, defaults to 0.02) — The maximum gradient norm.

8-bit LARS optimizer.

LARS32bit

class bitsandbytes.optim.LARS32bit

< >

( params lr momentum = 0 dampening = 0 weight_decay = 0 nesterov = False args = None min_8bit_size = 4096 percentile_clipping = 100 max_unorm = 0.02 )

__init__

< >

( params lr momentum = 0 dampening = 0 weight_decay = 0 nesterov = False args = None min_8bit_size = 4096 percentile_clipping = 100 max_unorm = 0.02 )

Parameters

  • params (torch.tensor) — The input parameters to optimize.
  • lr (float) — The learning rate.
  • momentum (float, defaults to 0) — The momentum value speeds up the optimizer by taking bigger steps.
  • dampening (float, defaults to 0) — The dampening value reduces the momentum of the optimizer.
  • weight_decay (float, defaults to 1e-2) — The weight decay value for the optimizer.
  • nesterov (bool, defaults to False) — Whether to use Nesterov momentum.
  • args (object, defaults to None) — An object with additional arguments.
  • min_8bit_size (int, defaults to 4096) — The minimum number of elements of the parameter tensors for 8-bit optimization.
  • percentile_clipping (int, defaults to 100) — Adapts clipping threshold automatically by tracking the last 100 gradient norms and clipping the gradient at a certain percentile to improve stability.
  • max_unorm (float, defaults to 0.02) — The maximum gradient norm.

32-bit LARS optimizer.

< > Update on GitHub

←LAMB Lion→