def get_args()

in modeling/utils/utils.py [0:0]

• 55 lines of code
• 1 McCabe index (conditional complexity)

def get_args():
	parser = argparse.ArgumentParser(description='')

	# general info
	parser.add_argument('--mode', type=str, required=True, help='')
	parser.add_argument('--task', type=str, required=True, help='which task [qr, coref, qr_coref] to perform? \
		`qr` for `qr-only` model; `coref` for `coref-only` model; `both` for `joint learning` model')
	parser.add_argument('--seed', type=int, default=1122)
	parser.add_argument('--model_name', type=str, required=True, help='model name, can be random but has to be provided')
	parser.add_argument('--model_name_or_path', type=str, default='gpt2')
	parser.add_argument('--checkpoint', type=str, default='', help='path of your model to save/load')
	parser.add_argument('--disable_display', type=str2bool, default=False, help='display progress bar or not')

	# data path
	parser.add_argument('--train_file', type=str, default='')
	parser.add_argument('--dev_file', type=str, default='')
	parser.add_argument('--test_file', type=str, default='')
	parser.add_argument('--dev_conll', type=str, default='')
	parser.add_argument('--test_conll', type=str, default='')

	# training
	parser.add_argument('--load_pretrained_weight', type=str2bool, default=True, \
		help='whether to load pretrained gpt2 weight or train from scratch')
	parser.add_argument('--train_batch_size', type=int, default=15, help='batch size of training per gpu')
	parser.add_argument('--eval_batch_size', type=int, default=1, help='batch size of evaluation per gpu')
	parser.add_argument('--gradient_accumulation_steps', type=int, default=4, help="steps for accumulating gradients")
	parser.add_argument('--learning_rate', type=float, default=6.25e-5)
	parser.add_argument('--adam_epsilon', type=float, default=1e-12)
	parser.add_argument('--max_grad_norm', type=float, default=1.0)
	parser.add_argument('--max_epoch', type=int, default=16)
	parser.add_argument('--use_scheduler', type=str2bool, default=True, help='whether to use lr scheduler')
	parser.add_argument('--warmup_steps', type=int, default=0)
	parser.add_argument('--train_size', type=int, default=-1, help='examples used for training. -1 means all data')
	parser.add_argument('--eval_interval', type=int, default=16000, help='how frequent (in steps) to evaluate the model during training')
	parser.add_argument('--no_improve_max', type=int, default=5, help='The max tolerance for model not improving')
	parser.add_argument('--eps', type=float, default=1e-12)
#	parser.add_argument('--fp16', type=str2bool, default=False, help='Whether to use float16')
#	parser.add_argument('--local_rank', type=int, default=-1, help='Local rank for distributed training (-1: not distributed)')

	# coreference resolution
	parser.add_argument('--coref_layer_idx', type=str2list, default=[10,11], \
		help='which layer to use for coref prediction, e.g., "1,5,11". 0<=n<=11')
	parser.add_argument('--n_coref_head', type=int, default=1, \
		help='How many heads to be used for coref prediction in each layer. 1<=n<=12')

	# coref2qr attention
	parser.add_argument('--use_coref_attn', type=str2bool, default=True, help="whether to use coref2qr attention")
	parser.add_argument('--coref_attn_layer', type=int, default=1, help="how many layer involved in coref2qr attention")
	parser.add_argument('--coref_attn_mention', type=str2bool, default=False, \
		help="whether to consider mentions' hidden states for coref2qr")
	parser.add_argument('--coref_attn_share_between_layer', type=str2bool, default=True, \
		help="whether to share parameters in coref2qr attention across layers")
	
	# binary classification
	parser.add_argument('--use_binary_cls', type=str2bool, default=True, help="whether to use binary classification")
	parser.add_argument('--filter_not_rewrite_loss', type=str2bool, default=True, \
		help="if True, lm loss of examples not requiring rewrite won't be considered")
	parser.add_argument('--copy_not_rewrite', type=str2bool, default=True, \
		help="if True, the model copies the input query as output when it predicts `no-rewrite`")
	parser.add_argument('--class0_loss_w', type=float, default=1., help="loss weight for `no-rewrite` class")
	parser.add_argument('--class1_loss_w', type=float, default=1.5, help="loss weight for `rewrite` class")
	
	# decoding
	parser.add_argument('--dec_max_len', type=int, default=100)
	parser.add_argument('--num_beams', type=int, default=1)
	parser.add_argument('--temperature', type=float, default=1.0)
	parser.add_argument('--decode_file', type=str, default='')

	args = parser.parse_args()
	print(args)
	verify_args(args)
	return args