train Namespace Reference

Functions
def	parse_args (parser)
def	reduce_tensor (tensor, num_gpus)
def	init_distributed (args, world_size, rank, group_name)
def	save_checkpoint (model, optimizer, epoch, config, amp_run, output_dir, model_name, local_rank, world_size)
def	get_last_checkpoint_filename (output_dir, model_name)
def	load_checkpoint (model, optimizer, epoch, config, amp_run, filepath, local_rank)
def	evaluating (model)
def	validate (model, criterion, valset, epoch, batch_iter, batch_size, world_size, collate_fn, distributed_run, rank, batch_to_gpu)
def	adjust_learning_rate (iteration, epoch, optimizer, learning_rate, anneal_steps, anneal_factor, rank)
def	main ()

Function Documentation

parse_args()

def train.parse_args

(

parser

)

Parse commandline arguments.

Here is the caller graph for this function:

reduce_tensor()

def train.reduce_tensor	(		tensor,
			num_gpus
	)

Here is the caller graph for this function:

init_distributed()

def train.init_distributed	(		args,
			world_size,
			rank,
			group_name
	)

Here is the caller graph for this function:

save_checkpoint()

def train.save_checkpoint	(		model,
			optimizer,
			epoch,
			config,
			amp_run,
			output_dir,
			model_name,
			local_rank,
			world_size
	)

Here is the caller graph for this function:

get_last_checkpoint_filename()

def train.get_last_checkpoint_filename	(		output_dir,
			model_name
	)

Here is the caller graph for this function:

load_checkpoint()

def train.load_checkpoint	(		model,
			optimizer,
			epoch,
			config,
			amp_run,
			filepath,
			local_rank
	)

Here is the caller graph for this function:

evaluating()

def train.evaluating

(

model

)

Temporarily switch to evaluation mode.

Here is the caller graph for this function:

validate()

def train.validate	(		model,
			criterion,
			valset,
			epoch,
			batch_iter,
			batch_size,
			world_size,
			collate_fn,
			distributed_run,
			rank,
			batch_to_gpu
	)

Handles all the validation scoring and printing

Here is the call graph for this function:

Here is the caller graph for this function:

adjust_learning_rate()

def train.adjust_learning_rate	(		iteration,
			epoch,
			optimizer,
			learning_rate,
			anneal_steps,
			anneal_factor,
			rank
	)

Here is the caller graph for this function:

main()

def train.main

(

)

Here is the call graph for this function: