def evaluate(self, eval_loader):
mAP, aps, pr_curves = eval_fcos_det(self.cfg,
self.criterion,
eval_loader,
self.model,
rescale=None)
logging.info(f"score_threshold:{self.cfg.TEST.SCORE_THRESHOLD}, "
f"nms_iou:{self.cfg.TEST.NMS_THRESHOLD}")
logging.info(f'mean Average Precision Across All Classes:{mAP}')
if mAP > self.best_performance or True:
# save PR curve
if self.cfg.TEST.PR_CURVE:
for class_idx, pr_curve in enumerate(pr_curves):
self.writer.add_figure(f'PR_curve/class_{class_idx + 1}',
pr_curve, self.iter)
logging.info('PR curves saved')
# save detected result
if self.cfg.TEST.IMAGE_DIR != '':
transform = PredictionTransform(resize=cfg.TEST.TEST_SIZE)
run_fcos_det_example(self.cfg, self.criterion,
self.cfg.TEST.IMAGE_DIR, transform,
self.model)
logging.info('test detected result saved')
return mAP
emb = InceptionResnetV1(pretrained='vggface2', classify=False)
assert cfg.MODEL.APPEARANCE.WEIGHTS != ''
load_eval_model(cfg.MODEL.APPEARANCE.WEIGHTS, emb)
emb.cuda().eval()
# read test frame images
if osp.isdir(args.frame_dir):
file_names = os.listdir(args.frame_dir)
jpg_paths = [
osp.join(args.frame_dir, file_name) for file_name in file_names
]
else:
raise FileNotFoundError
criterion = FCOSLoss(cfg, cfg.DATASET.NUM_CLASSES)
det_transform = PredictionTransform(size=cfg.TEST.TEST_SIZE)
ap_transform = TNTTransform(size=(cfg.TEST.CROP_SZIE))
crop_transform = FacenetInferenceTransform(size=(cfg.TEST.CROP_SZIE))
# get det result dict using the detector
det_result = {}
crop_im = {}
class_dict = {
0: 'DontCare',
1: 'Pedestrian',
2: 'Car',
3: 'Cyclist',
4: 'Van',
5: 'Truck',
6: 'Person',
7: 'Tram',
model.cuda()
model.eval()
if cfg.MODEL.RESUME_PATH != '':
if osp.exists(cfg.MODEL.RESUME_PATH):
print(f'==> model load from {cfg.MODEL.RESUME_PATH}')
checkpoint = torch.load(cfg.MODEL.RESUME_PATH)
if 'state_dict' in checkpoint:
model.load_state_dict(checkpoint['state_dict'])
else:
model.load_state_dict(checkpoint)
else:
print(f"==> checkpoint do not exists: \"{cfg.MODEL.RESUME_PATH}\"")
raise FileNotFoundError
transform = PredictionTransform(size=(352, 1216))
if not osp.exists('crop_det'):
os.mkdir('crop_det')
video_list = os.listdir(args.videos_root)
for video_name in video_list:
video_crop_path = osp.join('crop_det', video_name)
if not osp.exists(video_crop_path):
os.mkdir(video_crop_path)
video_path = osp.join(args.videos_root, video_name)
jpg_paths = []
# read test images
if osp.isfile(video_path):
jpg_paths = [video_path]
elif osp.isdir(video_path):