def test_net_on_dataset(args,
dataset_name,
proposal_file,
output_dir,
multi_gpu=False,
gpu_id=0):
"""Run inference on a dataset."""
dataset = JsonDatasetRel(dataset_name)
test_timer = Timer()
test_timer.tic()
if multi_gpu:
num_images = len(dataset.get_roidb(gt=args.do_val))
all_results = multi_gpu_test_net_on_dataset(args, dataset_name,
proposal_file, num_images,
output_dir)
else:
all_results = test_net(args,
dataset_name,
proposal_file,
output_dir,
gpu_id=gpu_id)
test_timer.toc()
logger.info('Total inference time: {:.3f}s'.format(
test_timer.average_time))
logger.info('Starting evaluation now...')
if dataset_name.find('vg') >= 0 or dataset_name.find('vrd') >= 0:
task_evaluation_vg_and_vrd.eval_rel_results(all_results, output_dir,
args.do_val)
else:
task_evaluation_sg.eval_rel_results(all_results, output_dir,
args.do_val, args.do_vis,
args.do_special)
return all_results
def get_metrics_det_boxes(model, timers, dataset_name):
# assert len(cfg.TEST.DATASETS) == 1
# dataset_name, proposal_file = get_inference_dataset(0, cfg.TEST.DATASETS)
# dataset = JsonDatasetRel(dataset_name)
model.eval()
roidb, dataset, start_ind, end_ind, total_num_images = get_roidb_and_dataset(
dataset_name, None, None, True)
num_images = len(roidb)
all_results = [None for _ in range(num_images)]
for i, entry in enumerate(roidb):
box_proposals = None
im = cv2.imread(entry['image'])
im_results = im_detect_rels(model, im, dataset_name, box_proposals,
False, timers)
im_results.update(dict(image=entry['image']))
# add gt
im_results.update(
dict(gt_sbj_boxes=entry['sbj_gt_boxes'],
gt_sbj_labels=entry['sbj_gt_classes'],
gt_obj_boxes=entry['obj_gt_boxes'],
gt_obj_labels=entry['obj_gt_classes'],
gt_prd_labels=entry['prd_gt_classes']))
all_results[i] = im_results
if dataset_name.find('vg') >= 0 or dataset_name.find(
'vrd') >= 0: # TODO: use pred_det to val
metrics = task_evaluation_vg_and_vrd.eval_rel_results(
all_results, None, True)
else:
metrics = task_evaluation_sg.eval_rel_results(all_results, None, True)
return metrics
if args.use_gt_boxes:
if args.use_gt_labels:
det_file = os.path.join(
args.output_dir, 'rel_detections_gt_boxes_prdcls_' +
args.load_ckpt.split('_')[-1].split('.')[-2] + '.pkl')
else:
det_file = os.path.join(
args.output_dir, 'rel_detections_gt_boxes_sgcls_' +
args.load_ckpt.split('_')[-1].split('.')[-2] + '.pkl')
else:
det_file = os.path.join(
args.output_dir, 'rel_detections_' +
args.load_ckpt.split('_')[-1].split('.')[-2] + '.pkl')
if os.path.exists(det_file):
logger.info('Loading results from {}'.format(det_file))
with open(det_file, 'rb') as f:
all_results = pickle.load(f)
logger.info('Starting evaluation now...')
if args.dataset.find('vg') >= 0 or args.dataset.find('vrd') >= 0:
task_evaluation_vg_and_vrd.eval_rel_results(
all_results, args.output_dir, args.do_val)
else:
task_evaluation_sg.eval_rel_results(all_results, args.output_dir,
args.do_val, args.do_vis,
args.do_special)
else:
run_inference(args,
ind_range=args.range,
multi_gpu_testing=args.multi_gpu_testing,
check_expected_results=True)
