bitsandbytes/pr_2036/en/reference/optim/lion.md

Lion

Lion (Evolved Sign Momentum) is a unique optimizer that uses the sign of the gradient to determine the update direction of the momentum. This makes Lion more memory-efficient and faster than AdamW which tracks and store the first and second-order moments.

Lion[[api-class]][[bitsandbytes.optim.Lion]]

bitsandbytes.optim.Lion[[bitsandbytes.optim.Lion]]

bitsandbytes.optim.Lion(params, lr = 0.0001, betas = (0.9, 0.99), weight_decay = 0, optim_bits = 32, args = None, min_8bit_size = 4096, is_paged = False)

Source

init[[bitsandbytes.optim.Lion.init]]

__init__(params, lr = 0.0001, betas = (0.9, 0.99), weight_decay = 0, optim_bits = 32, args = None, min_8bit_size = 4096, is_paged = False)

Source

Parameters:

params (torch.tensor) : The input parameters to optimize.

lr (float, defaults to 1e-4) : The learning rate.

betas (tuple(float, float), defaults to (0.9, 0.999)) : The beta values are the decay rates of the first and second-order moment of the optimizer.

weight_decay (float, defaults to 0) : The weight decay value for the optimizer.

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.

is_paged (bool, defaults to False) : Whether the optimizer is a paged optimizer or not.

Base Lion optimizer.

Lion8bit[[bitsandbytes.optim.Lion8bit]]

bitsandbytes.optim.Lion8bit[[bitsandbytes.optim.Lion8bit]]

bitsandbytes.optim.Lion8bit(params, lr = 0.0001, betas = (0.9, 0.99), weight_decay = 0, args = None, min_8bit_size = 4096, is_paged = False)

Source

init[[bitsandbytes.optim.Lion8bit.init]]

__init__(params, lr = 0.0001, betas = (0.9, 0.99), weight_decay = 0, args = None, min_8bit_size = 4096, is_paged = False)

Source

Parameters:

params (torch.tensor) : The input parameters to optimize.

lr (float, defaults to 1e-4) : The learning rate.

betas (tuple(float, float), defaults to (0.9, 0.999)) : The beta values are the decay rates of the first and second-order moment of the optimizer.

weight_decay (float, defaults to 0) : The weight decay value for the optimizer.

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.

is_paged (bool, defaults to False) : Whether the optimizer is a paged optimizer or not.

8-bit Lion optimizer.

Lion32bit[[bitsandbytes.optim.Lion32bit]]

bitsandbytes.optim.Lion32bit[[bitsandbytes.optim.Lion32bit]]

bitsandbytes.optim.Lion32bit(params, lr = 0.0001, betas = (0.9, 0.99), weight_decay = 0, args = None, min_8bit_size = 4096, is_paged = False)

Source

init[[bitsandbytes.optim.Lion32bit.init]]

__init__(params, lr = 0.0001, betas = (0.9, 0.99), weight_decay = 0, args = None, min_8bit_size = 4096, is_paged = False)

Source

Parameters:

params (torch.tensor) : The input parameters to optimize.

lr (float, defaults to 1e-4) : The learning rate.

betas (tuple(float, float), defaults to (0.9, 0.999)) : The beta values are the decay rates of the first and second-order moment of the optimizer.

weight_decay (float, defaults to 0) : The weight decay value for the optimizer.

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.

is_paged (bool, defaults to False) : Whether the optimizer is a paged optimizer or not.

32-bit Lion optimizer.

PagedLion[[bitsandbytes.optim.PagedLion]]

bitsandbytes.optim.PagedLion[[bitsandbytes.optim.PagedLion]]

bitsandbytes.optim.PagedLion(params, lr = 0.0001, betas = (0.9, 0.99), weight_decay = 0, optim_bits = 32, args = None, min_8bit_size = 4096)

Source

init[[bitsandbytes.optim.PagedLion.init]]

__init__(params, lr = 0.0001, betas = (0.9, 0.99), weight_decay = 0, optim_bits = 32, args = None, min_8bit_size = 4096)

Source

Parameters:

params (torch.tensor) : The input parameters to optimize.

lr (float, defaults to 1e-4) : The learning rate.

betas (tuple(float, float), defaults to (0.9, 0.999)) : The beta values are the decay rates of the first and second-order moment of the optimizer.

weight_decay (float, defaults to 0) : The weight decay value for the optimizer.

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.

Paged Lion optimizer.

PagedLion8bit[[bitsandbytes.optim.PagedLion8bit]]

bitsandbytes.optim.PagedLion8bit[[bitsandbytes.optim.PagedLion8bit]]

bitsandbytes.optim.PagedLion8bit(params, lr = 0.0001, betas = (0.9, 0.99), weight_decay = 0, args = None, min_8bit_size = 4096)

Source

init[[bitsandbytes.optim.PagedLion8bit.init]]

__init__(params, lr = 0.0001, betas = (0.9, 0.99), weight_decay = 0, args = None, min_8bit_size = 4096)

Source

Parameters:

params (torch.tensor) : The input parameters to optimize.

lr (float, defaults to 1e-4) : The learning rate.

betas (tuple(float, float), defaults to (0.9, 0.999)) : The beta values are the decay rates of the first and second-order moment of the optimizer.

weight_decay (float, defaults to 0) : The weight decay value for the optimizer.

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.

Paged 8-bit Lion optimizer.

PagedLion32bit[[bitsandbytes.optim.PagedLion32bit]]

bitsandbytes.optim.PagedLion32bit[[bitsandbytes.optim.PagedLion32bit]]

bitsandbytes.optim.PagedLion32bit(params, lr = 0.0001, betas = (0.9, 0.99), weight_decay = 0, args = None, min_8bit_size = 4096)

Source

init[[bitsandbytes.optim.PagedLion32bit.init]]

__init__(params, lr = 0.0001, betas = (0.9, 0.99), weight_decay = 0, args = None, min_8bit_size = 4096)

Source

Parameters:

params (torch.tensor) : The input parameters to optimize.

lr (float, defaults to 1e-4) : The learning rate.

betas (tuple(float, float), defaults to (0.9, 0.999)) : The beta values are the decay rates of the first and second-order moment of the optimizer.

weight_decay (float, defaults to 0) : The weight decay value for the optimizer.

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.

Paged 32-bit Lion optimizer.