def inference_single(net, path, fname, device):
image, im_info = get_image(path, fname)
np.set_printoptions(precision=2, suppress=True)
net.eval()
pred_boxes = net(image, im_info)
if if_set_nms:
from set_nms_utils import set_cpu_nms
n = pred_boxes.shape[0] // 2
idents = np.tile(np.arange(n)[:, None], (1, 2)).reshape(-1, 1)
pred_boxes = np.hstack((pred_boxes, idents))
keep = pred_boxes[:, -2] > 0.05
pred_boxes = pred_boxes[keep]
keep = set_cpu_nms(pred_boxes, 0.5)
pred_boxes = pred_boxes[keep]
else:
import det_tools_cuda as dtc
nms = dtc.nms
keep = nms(pred_boxes[:, :4], pred_boxes[:, 4], 0.5)
pred_boxes = pred_boxes[keep]
pred_boxes = np.array(pred_boxes)
keep = pred_boxes[:, -1] > 0.05
pred_boxes = pred_boxes[keep]
result_dict = dict(fname=fname,
height=int(im_info[0, -2]),
width=int(im_info[0, -1]),
dtboxes=boxes_dump(pred_boxes, False))
#rois=misc_utils.boxes_dump(rois[:, 1:], True))
return result_dict
def inference(model_file, device, records, result_queue):
torch.set_default_tensor_type('torch.FloatTensor')
net = network.Network()
net.cuda(device)
check_point = torch.load(model_file)
net.load_state_dict(check_point['state_dict'])
for record in records:
np.set_printoptions(precision=2, suppress=True)
net.eval()
image, gt_boxes, im_info, ID = get_data(record, device)
pred_boxes = net(image, im_info)
if if_set_nms:
from set_nms_utils import set_cpu_nms
n = pred_boxes.shape[0] // 2
idents = np.tile(np.arange(n)[:, None], (1, 2)).reshape(-1, 1)
pred_boxes = np.hstack((pred_boxes, idents))
keep = pred_boxes[:, -2] > 0.05
pred_boxes = pred_boxes[keep]
keep = set_cpu_nms(pred_boxes, 0.5)
pred_boxes = pred_boxes[keep]
else:
import det_tools_cuda as dtc
nms = dtc.nms
keep = nms(pred_boxes[:, :4], pred_boxes[:, 4], 0.5)
pred_boxes = pred_boxes[keep]
pred_boxes = np.array(pred_boxes)
keep = pred_boxes[:, -1] > 0.05
pred_boxes = pred_boxes[keep]
result_dict = dict(ID=ID,
height=int(im_info[0, -2]),
width=int(im_info[0, -1]),
dtboxes=boxes_dump(pred_boxes, False),
gtboxes=boxes_dump(gt_boxes, True))
#rois=misc_utils.boxes_dump(rois[:, 1:], True))
result_queue.put_nowait(result_dict)
def __inference(image, im_info, net):
pred_boxes = net(image, im_info)
n = pred_boxes.shape[0] // 2
idents = np.tile(np.arange(n)[:, None], (1, 2)).reshape(-1, 1)
pred_boxes = np.hstack((pred_boxes, idents))
keep = pred_boxes[:, -2] > 0.05
pred_boxes = pred_boxes[keep]
keep = set_cpu_nms(pred_boxes, 0.5)
pred_boxes = pred_boxes[keep]
return pred_boxes
def inference(args):
@jit.trace(symbolic=False)
def val_func():
pred_boxes = net(net.inputs)
return pred_boxes
# model path
saveDir = config.model_dir
evalDir = config.eval_dir
misc_utils.ensure_dir(evalDir)
model_file = os.path.join(saveDir,
'epoch_{}.pkl'.format(args.resume_weights))
assert os.path.exists(model_file)
# load model
net = network.Network()
net.eval()
check_point = mge.load(model_file)
net.load_state_dict(check_point['state_dict'])
ori_image, image, im_info = get_data(args.img_path)
net.inputs["image"].set_value(image.astype(np.float32))
net.inputs["im_info"].set_value(im_info)
pred_boxes = val_func().numpy()
num_tag = config.num_classes - 1
target_shape = (pred_boxes.shape[0] // num_tag // top_k, top_k)
pred_tags = (np.arange(num_tag) + 1).reshape(-1, 1)
pred_tags = np.tile(pred_tags, target_shape).reshape(-1, 1)
# nms
if if_set_nms:
from set_nms_utils import set_cpu_nms
n = pred_boxes.shape[0] // top_k
idents = np.tile(np.arange(n)[:, None], (1, top_k)).reshape(-1, 1)
pred_boxes = np.hstack((pred_boxes, idents))
keep = pred_boxes[:, -2] > args.thresh
pred_boxes = pred_boxes[keep]
pred_tags = pred_tags[keep]
keep = set_cpu_nms(pred_boxes, 0.5)
pred_boxes = pred_boxes[keep][:, :-1]
pred_tags = pred_tags[keep]
else:
from set_nms_utils import cpu_nms
keep = pred_boxes[:, -1] > args.thresh
pred_boxes = pred_boxes[keep]
pred_tags = pred_tags[keep]
keep = cpu_nms(pred_boxes, 0.5)
pred_boxes = pred_boxes[keep]
pred_tags = pred_tags[keep]
pred_tags = pred_tags.astype(np.int32).flatten()
pred_tags_name = np.array(config.class_names)[pred_tags]
visual_utils.draw_boxes(ori_image, pred_boxes[:, :-1], pred_boxes[:, -1],
pred_tags_name)
name = args.img_path.split('/')[-1].split('.')[-2]
fpath = '/data/jupyter/{}.png'.format(name)
cv2.imwrite(fpath, ori_image)
def inference(args):
@jit.trace(symbolic=False)
def val_func():
pred_boxes = net(net.inputs)
return pred_boxes
# model path
saveDir = config.model_dir
evalDir = config.eval_dir
misc_utils.ensure_dir(evalDir)
model_file = os.path.join(saveDir,
'epoch_{}.pkl'.format(args.resume_weights))
assert os.path.exists(model_file)
# load model
net = network.Network()
net.eval()
check_point = mge.load(model_file)
net.load_state_dict(check_point['state_dict'])
image, im_info = get_data(args.img_path)
net.inputs["image"].set_value(image.astype(np.float32))
net.inputs["im_info"].set_value(im_info)
pred_boxes = val_func().numpy()
num_tag = config.num_classes - 1
target_shape = (pred_boxes.shape[0] // num_tag // top_k, top_k)
pred_tags = (np.arange(num_tag) + 1).reshape(-1, 1)
pred_tags = np.tile(pred_tags, target_shape).reshape(-1, 1)
# nms
if if_set_nms:
from set_nms_utils import set_cpu_nms
n = pred_boxes.shape[0] // top_k
idents = np.tile(np.arange(n)[:, None], (1, top_k)).reshape(-1, 1)
pred_boxes = np.hstack((pred_boxes, idents))
keep = pred_boxes[:, -2] > 0.05
pred_boxes = pred_boxes[keep]
pred_tags = pred_tags[keep]
keep = set_cpu_nms(pred_boxes, 0.5)
pred_boxes = pred_boxes[keep][:, :-1]
pred_tags = pred_tags[keep].flatten()
else:
from set_nms_utils import cpu_nms
keep = pred_boxes[:, -1] > 0.05
pred_boxes = pred_boxes[keep]
pred_tags = pred_tags[keep]
keep = cpu_nms(pred_boxes, 0.5)
pred_boxes = pred_boxes[keep]
pred_tags = pred_tags[keep].flatten()
result_dict = dict(height=int(im_info[0, -2]),
width=int(im_info[0, -1]),
dtboxes=boxes_dump(pred_boxes, pred_tags))
name = args.img_path.split('/')[-1].split('.')[-2]
misc_utils.save_json_lines([result_dict], '{}.json'.format(name))
def inference_single_simple(image, net, category_id=0):
image, im_info = get_image_info(image)
np.set_printoptions(precision=2, suppress=True)
net.eval()
pred_boxes = net(image, im_info)
n = pred_boxes.shape[0] // 2
idents = np.tile(np.arange(n)[:, None], (1, 2)).reshape(-1, 1)
pred_boxes = np.hstack((pred_boxes, idents))
keep = pred_boxes[:, -2] > 0.05
pred_boxes = pred_boxes[keep]
keep = set_cpu_nms(pred_boxes, 0.5)
pred_boxes = pred_boxes[keep]
return boxes_dump(pred_boxes, False, category_id)
def inference(model_file, device, records, result_queue):
@jit.trace(symbolic=False)
def val_func():
pred_boxes = net(net.inputs)
return pred_boxes
net = network.Network()
net.eval()
check_point = mge.load(model_file)
net.load_state_dict(check_point['state_dict'])
for record in records:
np.set_printoptions(precision=2, suppress=True)
net.eval()
image, gt_boxes, im_info, ID = get_data(record, device)
net.inputs["image"].set_value(image.astype(np.float32))
net.inputs["im_info"].set_value(im_info)
pred_boxes = val_func().numpy()
num_tag = config.num_classes - 1
target_shape = (pred_boxes.shape[0] // num_tag // top_k, top_k)
pred_tags = (np.arange(num_tag) + 1).reshape(-1, 1)
pred_tags = np.tile(pred_tags, target_shape).reshape(-1, 1)
# nms
if if_set_nms:
from set_nms_utils import set_cpu_nms
n = pred_boxes.shape[0] // top_k
idents = np.tile(np.arange(n)[:, None], (1, top_k)).reshape(-1, 1)
pred_boxes = np.hstack((pred_boxes, idents))
keep = pred_boxes[:, -2] > 0.05
pred_boxes = pred_boxes[keep]
pred_tags = pred_tags[keep]
keep = set_cpu_nms(pred_boxes, 0.5)
pred_boxes = pred_boxes[keep][:, :-1]
pred_tags = pred_tags[keep].flatten()
else:
from set_nms_utils import cpu_nms
keep = pred_boxes[:, -1] > 0.05
pred_boxes = pred_boxes[keep]
pred_tags = pred_tags[keep]
keep = cpu_nms(pred_boxes, 0.5)
pred_boxes = pred_boxes[keep]
pred_tags = pred_tags[keep].flatten()
result_dict = dict(ID=ID,
height=int(im_info[0, -2]),
width=int(im_info[0, -1]),
dtboxes=boxes_dump(pred_boxes, pred_tags, False),
gtboxes=boxes_dump(gt_boxes, None, True))
result_queue.put_nowait(result_dict)
def inference_single(net, fname, img, device):
image, im_info = get_image_info(img)
np.set_printoptions(precision=2, suppress=True)
net.eval()
pred_boxes = net(image, im_info)
n = pred_boxes.shape[0] // 2
idents = np.tile(np.arange(n)[:, None], (1, 2)).reshape(-1, 1)
pred_boxes = np.hstack((pred_boxes, idents))
keep = pred_boxes[:, -2] > 0.05
pred_boxes = pred_boxes[keep]
keep = set_cpu_nms(pred_boxes, 0.5)
pred_boxes = pred_boxes[keep]
result_dict = dict(fname=fname,
height=int(im_info[0, -2]),
width=int(im_info[0, -1]),
dtboxes=boxes_dump(pred_boxes, False))
return result_dict