def visualizer(test_model, img):
# annotation load
d = Dataset()
annot = d.load_annot(cfg.testset)
gt_img_id = d.load_imgid(annot)
# human bbox load
if cfg.useGTbbox and cfg.testset in ['train', 'val']:
if cfg.testset == 'train':
dets = d.load_train_data(score=True)
else:
dets = d.load_val_data_with_annot()
dets.sort(key=lambda x: (x['image_id']))
else:
with open(cfg.human_det_path, 'r') as f:
dets = json.load(f)
dets = [i for i in dets if i['image_id'] in gt_img_id]
dets = [i for i in dets if i['category_id'] == 1]
dets = [i for i in dets if i['score'] > 0]
dets.sort(key=lambda x: (x['image_id'], x['score']), reverse=True)
img_id = []
for i in dets:
img_id.append(i['image_id'])
imgname = d.imgid_to_imgname(annot, img_id, cfg.testset)
for i in range(len(dets)):
dets[i]['imgpath'] = imgname[i]
dets_new = []
for det in dets:
if img in det['imgpath']:
dets_new.append(det)
break
dets = dets_new
# job assign (multi-gpu)
from tfflat.mp_utils import MultiProc
img_start = 0
ranges = [0]
img_num = len(np.unique([i['image_id'] for i in dets]))
images_per_gpu = int(img_num / len(args.gpu_ids.split(','))) + 1
for run_img in range(img_num):
img_end = img_start + 1
while img_end < len(dets) and dets[img_end]['image_id'] == dets[
img_start]['image_id']:
img_end += 1
if (run_img + 1) % images_per_gpu == 0 or (run_img + 1) == img_num:
ranges.append(img_end)
img_start = img_end
def func(gpu_id):
cfg.set_args(args.gpu_ids.split(',')[gpu_id])
tester = Tester(Model(), cfg)
tester.load_weights(test_model)
range = [ranges[gpu_id], ranges[gpu_id + 1]]
return test_net(tester, dets, range, gpu_id, True)
MultiGPUFunc = MultiProc(len(args.gpu_ids.split(',')), func)
result = MultiGPUFunc.work()
def test(test_model, logger):
import csv
anns = []
info = []
test_dir = '/home/lanhao/FashionAI/test3/'
test_ann_path = '/home/lanhao/FashionAI/test3/test_skirt.csv'
with open(test_ann_path, 'rb') as f:
reader = csv.reader(f)
for row in reader:
anns.append(row)
info.append(anns[0])
anns = anns[1:]
dets = []
for ann in anns:
imgname = ann[0]
imgname = test_dir + '/' + imgname
bbox = ann[2:]
imgpath = dict(imgpath=imgname, bbox=bbox)
dets.append(imgpath)
img_num = len(np.unique([i['imgpath'] for i in dets]))
from tfflat.mp_utils import MultiProc
img_start = 0
ranges = [0]
images_per_gpu = int(img_num / len(args.gpu_ids.split(','))) + 1
for run_img in range(img_num):
img_end = img_start + 1
while img_end < len(dets) and dets[img_end]['imgpath'] == dets[
img_start]['imgpath']:
img_end += 1
if (run_img + 1) % images_per_gpu == 0 or (run_img + 1) == img_num:
ranges.append(img_end)
img_start = img_end
def func(id):
cfg.set_args(args.gpu_ids.split(',')[id])
tester = Tester(Network(), cfg)
tester.load_weights(test_model)
range = [ranges[id], ranges[id + 1]]
return test_net(tester, logger, dets, range)
MultiGPUFunc = MultiProc(len(args.gpu_ids.split(',')), func)
all_res, dump_results = MultiGPUFunc.work()
# evaluation
results = []
results.append(info)
for i in range(len(anns)):
ann = anns[i]
keypoints = np.array(dump_results[i]['keypoints']).reshape(
(cfg.nr_skeleton, 3))
row = prepare_row(ann, keypoints)
results.append(row)
write_csv('result_skirt_test_0504.csv', results)
'''import json
def validate(test_model, logger):
eval_gt = COCO(cfg.gt_path)
#get validation
import json
with open(cfg.det_path, 'r') as f:
dets = json.load(f)["annotations"]
root_data_dir = "/home/data/COCO/MSCOCO/"
dets = [i for i in dets if i['image_id'] in eval_gt.imgs]
dets = [i for i in dets if i['category_id'] == 1]
dets.sort(key=lambda x: x['image_id'], reverse=True)
for i in dets:
i['imgpath'] = root_data_dir + 'val2017/' + str(
i['image_id']).zfill(12) + '.jpg'
i['score'] = 1.
img_num = len(np.unique([i['image_id'] for i in dets]))
from tfflat.mp_utils import MultiProc
img_start = 0
ranges = [0]
images_per_gpu = int(img_num / len(args.gpu_ids.split(','))) + 1
for run_img in range(img_num):
img_end = img_start + 1
while img_end < len(dets) and dets[img_end]['image_id'] == dets[
img_start]['image_id']:
img_end += 1
if (run_img + 1) % images_per_gpu == 0 or (run_img + 1) == img_num:
ranges.append(img_end)
img_start = img_end
def func(id):
cfg.set_args(args.gpu_ids.split(',')[id])
tester = Tester(Network(), cfg)
tester.load_weights(test_model)
range = [ranges[id], ranges[id + 1]]
return test_net(tester, logger, dets, range)
MultiGPUFunc = MultiProc(len(args.gpu_ids.split(',')), func)
all_res, dump_results = MultiGPUFunc.work()
# evaluation
result_path = osp.join(cfg.output_dir, 'results.json')
with open(result_path, 'w') as f:
json.dump(dump_results, f)
eval_dt = eval_gt.loadRes(result_path)
cocoEval = COCOeval(eval_gt, eval_dt, iouType='keypoints')
cocoEval.evaluate()
cocoEval.accumulate()
cocoEval.summarize()
def test(test_model):
# annotation load
d = Dataset()
annot = d.load_annot(cfg.testset)
# input pose load
#input_pose = d.coco_pose_load(annot)
#d.evaluation(input_pose, annot, cfg.result_dir, cfg.testset)
input_pose = d.input_pose_load(annot, cfg.testset)
# job assign (multi-gpu)
from tfflat.mp_utils import MultiProc
img_start = 0
ranges = [0]
img_num = len(np.unique([i['image_id'] for i in input_pose]))
print(img_num)
images_per_gpu = int(img_num / len(args.gpu_ids.split(','))) + 1
for run_img in range(img_num):
img_end = img_start + 1
while img_end < len(input_pose) and input_pose[img_end][
'image_id'] == input_pose[img_start]['image_id']:
img_end += 1
if (run_img + 1) % images_per_gpu == 0 or (run_img + 1) == img_num:
ranges.append(img_end)
img_start = img_end
def func(gpu_id):
cfg.set_args(args.gpu_ids.split(',')[gpu_id])
tester = Tester(Model(), cfg)
tester.load_weights(test_model)
range = [ranges[gpu_id], ranges[gpu_id + 1]]
return test_net(tester, input_pose, range, gpu_id)
MultiGPUFunc = MultiProc(len(args.gpu_ids.split(',')), func)
result = MultiGPUFunc.work()
# func(0)
# evaluation
d.evaluation(result, annot, cfg.result_dir, cfg.testset)
def test(test_model, logger):
eval_gt = COCO(cfg.gt_path)
import json
with open(cfg.det_path, 'r') as f:
dets = json.load(f)
test_subset = False
if test_subset:
eval_gt.imgs = dict(list(eval_gt.imgs.items())[:100])
anns = dict()
for i in eval_gt.imgs:
for j in eval_gt.getAnnIds(i):
anns[j] = eval_gt.anns[j]
eval_gt.anns = anns
dets = [i for i in dets if i['image_id'] in eval_gt.imgs]
dets = [i for i in dets if i['category_id'] == 1]
dets.sort(key=lambda x: (x['image_id'], x['score']), reverse=True)
for i in dets:
i['imgpath'] = '/home/dx/data/coco/val2014/COCO_val2014_000000%06d.jpg' % i['image_id']
img_num = len(np.unique([i['image_id'] for i in dets]))
use_gtboxes = False
if use_gtboxes:
d = COCOJoints()
coco_train_data, coco_test_data = d.load_data()
coco_test_data.sort(key=lambda x: x['imgid'])
for i in coco_test_data:
i['image_id'] = i['imgid']
i['score'] = 1.
dets = coco_test_data
from tfflat.mp_utils import MultiProc
img_start = 0
ranges = [0]
images_per_gpu = int(img_num / len(args.gpu_ids.split(','))) + 1
for run_img in range(img_num):
img_end = img_start + 1
while img_end < len(dets) and dets[img_end]['image_id'] == dets[img_start]['image_id']:
img_end += 1
if (run_img + 1) % images_per_gpu == 0 or (run_img + 1) == img_num:
ranges.append(img_end)
img_start = img_end
def func(id):
cfg.set_args(args.gpu_ids.split(',')[id])
tester = Tester(Network(), cfg)
tester.load_weights(test_model)
range = [ranges[id], ranges[id + 1]]
return test_net(tester, logger, dets, range)
MultiGPUFunc = MultiProc(len(args.gpu_ids.split(',')), func)
all_res, dump_results = MultiGPUFunc.work()
# evaluation
result_path = osp.join(cfg.output_dir, 'results.json')
with open(result_path, 'w') as f:
json.dump(dump_results, f)
eval_dt = eval_gt.loadRes(result_path)
cocoEval = COCOeval(eval_gt, eval_dt, iouType='keypoints')
cocoEval.evaluate()
cocoEval.accumulate()
cocoEval.summarize()
def test(test_model):
# annotation load
d = Dataset()
d.setup_paths(cfg)
annot = d.load_annot(cfg.TEST.testset)
gt_img_id = d.load_imgid(annot)
if args.use_dets:
print("loading detections from human detector")
with open(cfg.human_det_path, 'r') as f:
dets = json.load(f)
dets = [i for i in dets if i['image_id'] in gt_img_id]
dets = [i for i in dets if i['category_id'] == 1]
dets = [i for i in dets if i['score'] > 0]
dets.sort(key=lambda x: (x['image_id'], x['score']), reverse=True)
img_id = []
for i in dets:
img_id.append(i['image_id'])
imgname = d.imgid_to_imgname(annot, img_id, cfg.TEST.testset)
for i in range(len(dets)):
dets[i]['imgpath'] = imgname[i]
else:
print("loading ground truth detections")
if cfg.TEST.testset == 'train':
dets = d.load_train_data(score=True)
else:
dets = d.load_val_data_with_annot(cfg.TEST.testset)
dets.sort(key=lambda x: (x['image_id']))
# job assign (multi-gpu)
from tfflat.mp_utils import MultiProc
img_start = 0
ranges = [0]
img_num = len(np.unique([i['image_id'] for i in dets]))
images_per_gpu = int(img_num / len(args.gpu_ids.split(','))) + 1
for run_img in range(img_num):
img_end = img_start + 1
while img_end < len(dets) and dets[img_end]['image_id'] == dets[
img_start]['image_id']:
img_end += 1
if (run_img + 1) % images_per_gpu == 0 or (run_img + 1) == img_num:
ranges.append(img_end)
img_start = img_end
def func(gpu_id):
config.set_args(args.gpu_ids.split(',')[gpu_id])
tester = Tester(Model(), cfg)
tester.load_weights(test_model)
range = [ranges[gpu_id], ranges[gpu_id + 1]]
if (cfg.MODEL.occluded_detection):
return test_net_occ(tester, dets, range, gpu_id, d.sigmas)
else:
return test_net(tester, dets, range, gpu_id, d.sigmas)
MultiGPUFunc = MultiProc(len(args.gpu_ids.split(',')), func)
result = MultiGPUFunc.work()
# evaluation
d.evaluation(result, annot, cfg.result_dir, cfg.TEST.testset,
cfg.EXPERIMENT_NAME, args.test_epoch)