training::config::TrainingConfig¶

Public Functions¶

	Name
dict	to_lora_params(self self)
dict	to_args_dict(self self)
"TrainingConfig"	from_yaml(cls cls, Path yaml_path, Optional specialist[str] =None)

Public Attributes¶

	Name
str	fine_tune_type
str	optimizer
int	batch_size
int	iters
int	val_batches
float	learning_rate
int	steps_per_report
int	steps_per_eval
int	save_every
int	num_layers
bool	grad_checkpoint
int	grad_accumulation_steps
bool	mask_prompt
Optional	report_to
Optional	project_name
int	seed
int	lora_rank
float	lora_dropout
float	lora_scale
bool	use_qlora
	lora_rank
	lora_dropout
	lora_scale
	fine_tune_type
	optimizer
	batch_size
	iters
	val_batches
	learning_rate
	steps_per_report
	steps_per_eval
	save_every
	num_layers
	grad_checkpoint
	grad_accumulation_steps
	mask_prompt
	report_to
	project_name
	seed

Public Functions Documentation¶

function to_lora_params¶

dict to_lora_params(
    self self
)

function to_args_dict¶

dict to_args_dict(
    self self
)

function from_yaml¶

"TrainingConfig" from_yaml(
    cls cls,
    Path yaml_path,
    Optional specialist[str] =None
)

Public Attributes Documentation¶

variable fine_tune_type¶

static str fine_tune_type =  "lora";

variable optimizer¶

static str optimizer =  "adamw";

variable batch_size¶

static int batch_size =  4;

variable iters¶

static int iters =  1000;

variable val_batches¶

static int val_batches =  25;

variable learning_rate¶

static float learning_rate =  1e-5;

variable steps_per_report¶

static int steps_per_report =  50;

variable steps_per_eval¶

static int steps_per_eval =  200;

variable save_every¶

static int save_every =  200;

variable num_layers¶

static int num_layers =  16;

variable grad_checkpoint¶

static bool grad_checkpoint =  True;

variable grad_accumulation_steps¶

static int grad_accumulation_steps =  1;

variable mask_prompt¶

static bool mask_prompt =  False;

variable report_to¶

static Optional report_to =  None;

variable project_name¶

static Optional project_name =  None;

variable seed¶

static int seed =  42;

variable lora_rank¶

static int lora_rank =  16;

variable lora_dropout¶

static float lora_dropout =  0.05;

variable lora_scale¶

static float lora_scale =  20.0;

variable use_qlora¶

static bool use_qlora =  True;

variable lora_rank¶

lora_rank;

variable lora_dropout¶

lora_dropout;

variable lora_scale¶

lora_scale;

variable fine_tune_type¶

fine_tune_type;

variable optimizer¶

optimizer;

variable batch_size¶

batch_size;

variable iters¶

iters;

variable val_batches¶

val_batches;

variable learning_rate¶

learning_rate;

variable steps_per_report¶

steps_per_report;

variable steps_per_eval¶

steps_per_eval;

variable save_every¶

save_every;

variable num_layers¶

num_layers;

variable grad_checkpoint¶

grad_checkpoint;

variable grad_accumulation_steps¶

grad_accumulation_steps;

variable mask_prompt¶

mask_prompt;

variable report_to¶

report_to;

variable project_name¶

project_name;

variable seed¶

seed;

Updated on 2026-07-25 at 22:56:57 +0000