dest='timelapse')
parser.add_argument(
'--rangewidth',
action='store',
default=-1,
help="uses range_width to limit number of images to match with.",
type=int,
dest='rangewidth')
__doc__ += '\n' + parser.format_help()
from options import Options
args = Options()
args.add_argument("work_megapix", 0.6)
args.add_argument('features', list(FEATURES_FIND_CHOICES.keys())[0])
args.add_argument('matcher', 'homography')
args.add_argument('estimator', list(ESTIMATOR_CHOICES.keys())[0])
args.add_argument('match_conf', 0.3)
args.add_argument('conf_thresh', 0.3)
args.add_argument('ba', list(BA_COST_CHOICES.keys())[0])
args.add_argument('ba_refine_mask', 'xxxxx')
args.add_argument('wave_correct', WAVE_CORRECT_CHOICES[0])
args.add_argument('save_graph', None)
args.add_argument('warp', WARP_CHOICES[3])
args.add_argument('seam_megapix', 0.1)
args.add_argument('seam', list(SEAM_FIND_CHOICES.keys())[0])
args.add_argument('compose_megapix', -1)
args.add_argument('expos_comp', list(EXPOS_COMP_CHOICES.keys())[0])
args.add_argument('expos_comp_nr_feeds', 1)
from utils import AverageMeter, Logger
from options import Options
from models.rram import get_rram_param_groups
from sparse_sgd import *
model_names = sorted(name for name in models.__dict__
if name.islower() and not name.startswith("__")
and callable(models.__dict__[name]))
options = Options(description='PyTorch ImageNet Sparse Training')
options.add_argument('--arch',
'-a',
metavar='ARCH',
default='resnet50',
choices=model_names,
help='model architecture: ' + ' | '.join(model_names) +
' (default: resnet50)')
options.set_defaults(data='gen/imagenet',
epochs=90,
batch_size=256,
lr=0.1,
momentum=0.9,
nesterov=False,
weight_decay=1e-4,
lr_decay=0.1,
lr_decay_step='30')
best_prec1 = 0
