def get_vgg16_test(system_dict):
from symnet.symbol_vgg import get_vgg_test
return get_vgg_test(anchor_scales=system_dict["rpn_anchor_scales"], anchor_ratios=system_dict["rpn_anchor_ratios"],
rpn_feature_stride=system_dict["rpn_feat_stride"], rpn_pre_topk=system_dict["rpn_pre_nms_topk"],
rpn_post_topk=system_dict["rpn_post_nms_topk"], rpn_nms_thresh=system_dict["rpn_nms_thresh"],
rpn_min_size=system_dict["rpn_min_size"],
num_classes=system_dict["rcnn_num_classes"], rcnn_feature_stride=system_dict["rcnn_feat_stride"],
rcnn_pooled_size=system_dict["rcnn_pooled_size"], rcnn_batch_size=system_dict["rcnn_batch_size"])
def get_vgg16_test(args):
from symnet.symbol_vgg import get_vgg_test
if not args.params:
args.params = 'model/vgg16-0010.params'
args.img_pixel_means = (123.68, 116.779, 103.939)
args.img_pixel_stds = (1.0, 1.0, 1.0)
args.net_fixed_params = ['conv1', 'conv2']
args.rpn_feat_stride = 16
args.rcnn_feat_stride = 16
args.rcnn_pooled_size = (7, 7)
return get_vgg_test(anchor_scales=args.rpn_anchor_scales, anchor_ratios=args.rpn_anchor_ratios,
rpn_feature_stride=args.rpn_feat_stride, rpn_pre_topk=args.rpn_pre_nms_topk,
rpn_post_topk=args.rpn_post_nms_topk, rpn_nms_thresh=args.rpn_nms_thresh,
rpn_min_size=args.rpn_min_size,
num_classes=args.rcnn_num_classes, rcnn_feature_stride=args.rcnn_feat_stride,
rcnn_pooled_size=args.rcnn_pooled_size, rcnn_batch_size=args.rcnn_batch_size)
def get_vgg16_test(args, config):
from symnet.symbol_vgg import get_vgg_test
if not args.params:
args.params = 'model/vgg16-0010.params'
# config = Config('configs/vgg_step_{}.yml'.format(args.step))
return get_vgg_test(anchor_scales=config.rpn['rpn_anchor_scales'],
anchor_ratios=config.rpn['rpn_anchor_ratios'],
rpn_feature_stride=config.rpn['rpn_feat_stride'],
rpn_pre_topk=config.rpn['rpn_pre_nms_topk'],
rpn_post_topk=config.rpn['rpn_post_nms_topk'],
rpn_nms_thresh=config.rpn['rpn_nms_thresh'],
rpn_min_size=config.rpn['rpn_min_size'],
num_classes=config.rcnn['rcnn_num_classes'],
rcnn_feature_stride=config.rcnn['rcnn_feat_stride'],
rcnn_pooled_size=config.rcnn['rcnn_pooled_size'],
rcnn_batch_size=args.rcnn_batch_size,
isBin=config.train_param['is_rcnn_top_bin'],
step=args.step)
def get_vgg16_test(args):
from symnet.symbol_vgg import get_vgg_test
if not args.params:
args.params = 'model/vgg16-0010.params'
args.rpn_feat_stride = 16
args.rcnn_feat_stride = 16
args.rcnn_pooled_size = (7, 7)
return get_vgg_test(anchor_scales=args.rpn_anchor_scales,
anchor_ratios=args.rpn_anchor_ratios,
rpn_feature_stride=args.rpn_feat_stride,
rpn_pre_topk=args.rpn_pre_nms_topk,
rpn_post_topk=args.rpn_post_nms_topk,
rpn_nms_thresh=args.rpn_nms_thresh,
rpn_min_size=args.rpn_min_size,
num_classes=args.rcnn_num_classes,
rcnn_feature_stride=args.rcnn_feat_stride,
rcnn_pooled_size=args.rcnn_pooled_size,
rcnn_batch_size=args.rcnn_batch_size,
isBin=args.is_bin_new,
step=args.step_new)
def get_vgg16_test(args):
from symnet.symbol_vgg import get_vgg_test
if not args.params:
args.params = 'model/vgg16-0010.params'
args.img_pixel_means = (123.68, 116.779, 103.939)
args.img_pixel_stds = (1.0, 1.0, 1.0)
args.net_fixed_params = ['conv1', 'conv2']
args.rpn_feat_stride = 16
args.rcnn_feat_stride = 16
args.rcnn_pooled_size = (7, 7)
return get_vgg_test(anchor_scales=args.rpn_anchor_scales,
anchor_ratios=args.rpn_anchor_ratios,
rpn_feature_stride=args.rpn_feat_stride,
rpn_pre_topk=args.rpn_pre_nms_topk,
rpn_post_topk=args.rpn_post_nms_topk,
rpn_nms_thresh=args.rpn_nms_thresh,
rpn_min_size=args.rpn_min_size,
num_classes=args.rcnn_num_classes,
rcnn_feature_stride=args.rcnn_feat_stride,
rcnn_pooled_size=args.rcnn_pooled_size,
rcnn_batch_size=args.rcnn_batch_size)
def get_vgg16_test(system_dict):
'''
Internal function: Select vgg16 params
Args:
system_dict (dict): Dictionary of all the parameters selected for training
Returns:
mxnet model: Vgg16 model
'''
from symnet.symbol_vgg import get_vgg_test
return get_vgg_test(anchor_scales=system_dict["rpn_anchor_scales"],
anchor_ratios=system_dict["rpn_anchor_ratios"],
rpn_feature_stride=system_dict["rpn_feat_stride"],
rpn_pre_topk=system_dict["rpn_pre_nms_topk"],
rpn_post_topk=system_dict["rpn_post_nms_topk"],
rpn_nms_thresh=system_dict["rpn_nms_thresh"],
rpn_min_size=system_dict["rpn_min_size"],
num_classes=system_dict["rcnn_num_classes"],
rcnn_feature_stride=system_dict["rcnn_feat_stride"],
rcnn_pooled_size=system_dict["rcnn_pooled_size"],
rcnn_batch_size=system_dict["rcnn_batch_size"])
