def load_network(self): # load weights weights_filename = osp.expanduser(self.weights) if osp.isfile(weights_filename): loc_func = lambda storage, loc: storage checkpoint = torch.load(weights_filename, map_location=loc_func) load_state_dict(self.eval_net, checkpoint['model_state_dict']) print 'Loaded weights from {:s}'.format(weights_filename) else: print 'Could not load weights from {:s}'.format(weights_filename) sys.exit(-1)
if (args.model.find('mapnet') >= 0) or args.pose_graph: model = MapNet(mapnet=posenet) else: model = posenet model.eval() # loss functions t_criterion = lambda t_pred, t_gt: np.linalg.norm(t_pred - t_gt) q_criterion = quaternion_angular_error # load weights weights_filename = osp.expanduser(args.weights) if osp.isfile(weights_filename): loc_func = lambda storage, loc: storage checkpoint = torch.load(weights_filename, map_location=loc_func) load_state_dict(model, checkpoint['model_state_dict']) print 'Loaded weights from {:s}'.format(weights_filename) else: print 'Could not load weights from {:s}'.format(weights_filename) sys.exit(-1) data_dir = osp.join('..', 'data', args.dataset) stats_filename = osp.join(data_dir, args.scene, 'stats.txt') stats = np.loadtxt(stats_filename) # transformer data_transform = transforms.Compose([ transforms.Resize(256), transforms.ToTensor(), transforms.Normalize(mean=stats[0], std=np.sqrt(stats[1])) ]) target_transform = transforms.Lambda(lambda x: torch.from_numpy(x).float())
calc_vos_safe, ) from dataset_loaders.composite import MF import argparse import os import os.path as osp import sys import numpy as np import matplotlib DISPLAY = "DISPLAY" in os.environ if not DISPLAY: matplotlib.use("Agg") import matplotlib.pyplot as plt from mpl_toolkits.mplot3d import Axes3D import configparser import torch.cuda from torch.utils.data import DataLoader from torchvision import transforms, models import cPickle dropout = 0.5 feature_extractor = models.resnet34(pretrained=False) model = PoseNet(feature_extractor, droprate=dropout, pretrained=False) weights_filename = "./weights.pth" loc_func = lambda storage, loc: storage checkpoint = torch.load(weights_filename, map_location=loc_func) load_state_dict(model, checkpoint["model_state_dict"]) print "Loaded weights from {:s}".format(weights_filename)