def test_revert_sync_batchnorm():
conv_syncbn = ConvModule(3, 8, 2, norm_cfg=dict(type='SyncBN')).to('cpu')
conv_syncbn.train()
x = torch.randn(1, 3, 10, 10)
# Will raise an ValueError saying SyncBN does not run on CPU
with pytest.raises(ValueError):
y = conv_syncbn(x)
conv_bn = revert_sync_batchnorm(conv_syncbn)
y = conv_bn(x)
assert y.shape == (1, 8, 9, 9)
assert conv_bn.training == conv_syncbn.training
conv_syncbn.eval()
conv_bn = revert_sync_batchnorm(conv_syncbn)
assert conv_bn.training == conv_syncbn.training
def test_psenet(cfg_file):
model = _get_detector_cfg(cfg_file)
model['pretrained'] = None
from mmocr.models import build_detector
detector = build_detector(model)
detector = revert_sync_batchnorm(detector)
input_shape = (1, 3, 224, 224)
num_kernels = 7
mm_inputs = _demo_mm_inputs(num_kernels, input_shape)
imgs = mm_inputs.pop('imgs')
img_metas = mm_inputs.pop('img_metas')
gt_kernels = mm_inputs.pop('gt_kernels')
gt_mask = mm_inputs.pop('gt_mask')
# Test forward train
losses = detector.forward(
imgs, img_metas, gt_kernels=gt_kernels, gt_mask=gt_mask)
assert isinstance(losses, dict)
# Test forward test
with torch.no_grad():
img_list = [g[None, :] for g in imgs]
batch_results = []
for one_img, one_meta in zip(img_list, img_metas):
result = detector.forward([one_img], [[one_meta]],
return_loss=False)
batch_results.append(result)
# Test show result
results = {'boundary_result': [[0, 0, 1, 0, 1, 1, 0, 1, 0.9]]}
img = np.random.rand(5, 5)
detector.show_result(img, results)
def build_model(config_file):
device = 'cpu'
model = init_detector(config_file, checkpoint=None, device=device)
model = revert_sync_batchnorm(model)
if model.cfg.data.test['type'] == 'ConcatDataset':
model.cfg.data.test.pipeline = model.cfg.data.test['datasets'][
0].pipeline
return model
def test_fcenet(cfg_file):
model = _get_detector_cfg(cfg_file)
model['pretrained'] = None
from mmocr.models import build_detector
detector = build_detector(model)
detector = revert_sync_batchnorm(detector)
detector = detector.cuda()
fourier_degree = 5
input_shape = (1, 3, 256, 256)
(n, c, h, w) = input_shape
imgs = torch.randn(n, c, h, w).float().cuda()
img_metas = [{
'img_shape': (h, w, c),
'ori_shape': (h, w, c),
'pad_shape': (h, w, c),
'filename': '<demo>.png',
'scale_factor': np.array([1, 1, 1, 1]),
'flip': False,
} for _ in range(n)]
p3_maps = []
p4_maps = []
p5_maps = []
for _ in range(n):
p3_maps.append(
np.random.random((5 + 4 * fourier_degree, h // 8, w // 8)))
p4_maps.append(
np.random.random((5 + 4 * fourier_degree, h // 16, w // 16)))
p5_maps.append(
np.random.random((5 + 4 * fourier_degree, h // 32, w // 32)))
# Test forward train
losses = detector.forward(imgs,
img_metas,
p3_maps=p3_maps,
p4_maps=p4_maps,
p5_maps=p5_maps)
assert isinstance(losses, dict)
# Test forward test
with torch.no_grad():
img_list = [g[None, :] for g in imgs]
batch_results = []
for one_img, one_meta in zip(img_list, img_metas):
result = detector.forward([one_img], [[one_meta]],
return_loss=False)
batch_results.append(result)
# Test show result
results = {'boundary_result': [[0, 0, 1, 0, 1, 1, 0, 1, 0.9]]}
img = np.random.rand(5, 5)
detector.show_result(img, results)
def test_textsnake(cfg_file):
model = _get_detector_cfg(cfg_file)
model['pretrained'] = None
from mmocr.models import build_detector
detector = build_detector(model)
detector = revert_sync_batchnorm(detector)
input_shape = (1, 3, 224, 224)
num_kernels = 1
mm_inputs = _demo_mm_inputs(num_kernels, input_shape)
imgs = mm_inputs.pop('imgs')
img_metas = mm_inputs.pop('img_metas')
gt_text_mask = mm_inputs.pop('gt_text_mask')
gt_center_region_mask = mm_inputs.pop('gt_center_region_mask')
gt_mask = mm_inputs.pop('gt_mask')
gt_radius_map = mm_inputs.pop('gt_radius_map')
gt_sin_map = mm_inputs.pop('gt_sin_map')
gt_cos_map = mm_inputs.pop('gt_cos_map')
# Test forward train
losses = detector.forward(
imgs,
img_metas,
gt_text_mask=gt_text_mask,
gt_center_region_mask=gt_center_region_mask,
gt_mask=gt_mask,
gt_radius_map=gt_radius_map,
gt_sin_map=gt_sin_map,
gt_cos_map=gt_cos_map)
assert isinstance(losses, dict)
# Test forward test get_boundary
maps = torch.zeros((1, 5, 224, 224), dtype=torch.float)
maps[:, 0:2, :, :] = -10.
maps[:, 0, 60:100, 12:212] = 10.
maps[:, 1, 70:90, 22:202] = 10.
maps[:, 2, 70:90, 22:202] = 0.
maps[:, 3, 70:90, 22:202] = 1.
maps[:, 4, 70:90, 22:202] = 10.
one_meta = img_metas[0]
result = detector.bbox_head.get_boundary(maps, [one_meta], False)
assert 'boundary_result' in result
assert 'filename' in result
# Test show result
results = {'boundary_result': [[0, 0, 1, 0, 1, 1, 0, 1, 0.9]]}
img = np.random.rand(5, 5)
detector.show_result(img, results)
def test_panet(cfg_file):
model = _get_detector_cfg(cfg_file)
model['pretrained'] = None
from mmocr.models import build_detector
detector = build_detector(model)
detector = revert_sync_batchnorm(detector)
input_shape = (1, 3, 224, 224)
num_kernels = 2
mm_inputs = _demo_mm_inputs(num_kernels, input_shape)
imgs = mm_inputs.pop('imgs')
img_metas = mm_inputs.pop('img_metas')
gt_kernels = mm_inputs.pop('gt_kernels')
gt_mask = mm_inputs.pop('gt_mask')
# Test forward train
losses = detector.forward(imgs,
img_metas,
gt_kernels=gt_kernels,
gt_mask=gt_mask)
assert isinstance(losses, dict)
# Test forward test
with torch.no_grad():
img_list = [g[None, :] for g in imgs]
batch_results = []
for one_img, one_meta in zip(img_list, img_metas):
result = detector.forward([one_img], [[one_meta]],
return_loss=False)
batch_results.append(result)
# Test onnx export
detector.forward = partial(detector.simple_test,
img_metas=img_metas,
rescale=True)
with tempfile.TemporaryDirectory() as tmpdirname:
onnx_path = f'{tmpdirname}/tmp.onnx'
torch.onnx.export(detector, (img_list[0], ),
onnx_path,
input_names=['input'],
output_names=['output'],
export_params=True,
keep_initializers_as_inputs=False)
# Test show result
results = {'boundary_result': [[0, 0, 1, 0, 1, 1, 0, 1, 0.9]]}
img = np.random.rand(5, 5)
detector.show_result(img, results)
def main():
args = parse_args()
assert (
args.out or args.eval or args.format_only or args.show
or args.show_dir), (
'Please specify at least one operation (save/eval/format/show the '
'results / save the results) with the argument "--out", "--eval"'
', "--format-only", "--show" or "--show-dir".')
if args.eval and args.format_only:
raise ValueError('--eval and --format_only cannot be both specified.')
if args.out is not None and not args.out.endswith(('.pkl', '.pickle')):
raise ValueError('The output file must be a pkl file.')
cfg = Config.fromfile(args.config)
if args.cfg_options is not None:
cfg.merge_from_dict(args.cfg_options)
# import modules from string list.
if cfg.get('custom_imports', None):
from mmcv.utils import import_modules_from_strings
import_modules_from_strings(**cfg['custom_imports'])
# set cudnn_benchmark
if cfg.get('cudnn_benchmark', False):
torch.backends.cudnn.benchmark = True
if cfg.model.get('pretrained'):
cfg.model.pretrained = None
if cfg.model.get('neck'):
if isinstance(cfg.model.neck, list):
for neck_cfg in cfg.model.neck:
if neck_cfg.get('rfp_backbone'):
if neck_cfg.rfp_backbone.get('pretrained'):
neck_cfg.rfp_backbone.pretrained = None
elif cfg.model.neck.get('rfp_backbone'):
if cfg.model.neck.rfp_backbone.get('pretrained'):
cfg.model.neck.rfp_backbone.pretrained = None
# in case the test dataset is concatenated
samples_per_gpu = 1
if isinstance(cfg.data.test, dict):
samples_per_gpu = (cfg.data.get('test_dataloader', {})).get(
'samples_per_gpu', cfg.data.get('samples_per_gpu', 1))
if samples_per_gpu > 1:
# Support batch_size > 1 in test for text recognition
# by disable MultiRotateAugOCR since it is useless for most case
cfg = disable_text_recog_aug_test(cfg)
if cfg.data.test.get('pipeline', None) is not None:
# Replace 'ImageToTensor' to 'DefaultFormatBundle'
cfg.data.test.pipeline = replace_ImageToTensor(
cfg.data.test.pipeline)
elif isinstance(cfg.data.test, list):
for ds_cfg in cfg.data.test:
ds_cfg.test_mode = True
samples_per_gpu = max(
[ds_cfg.pop('samples_per_gpu', 1) for ds_cfg in cfg.data.test])
if samples_per_gpu > 1:
for ds_cfg in cfg.data.test:
ds_cfg.pipeline = replace_ImageToTensor(ds_cfg.pipeline)
# init distributed env first, since logger depends on the dist info.
if args.launcher == 'none':
distributed = False
else:
distributed = True
init_dist(args.launcher, **cfg.dist_params)
# build the dataloader
dataset = build_dataset(cfg.data.test, dict(test_mode=True))
# step 1: give default values and override (if exist) from cfg.data
loader_cfg = {
**dict(seed=cfg.get('seed'), drop_last=False, dist=distributed),
**({} if torch.__version__ != 'parrots' else dict(
prefetch_num=2,
pin_memory=False,
)),
**dict((k, cfg.data[k]) for k in [
'workers_per_gpu',
'seed',
'prefetch_num',
'pin_memory',
'persistent_workers',
] if k in cfg.data)
}
test_loader_cfg = {
**loader_cfg,
**dict(shuffle=False, drop_last=False),
**cfg.data.get('test_dataloader', {}),
**dict(samples_per_gpu=samples_per_gpu)
}
data_loader = build_dataloader(dataset, **test_loader_cfg)
# build the model and load checkpoint
cfg.model.train_cfg = None
model = build_detector(cfg.model, test_cfg=cfg.get('test_cfg'))
model = revert_sync_batchnorm(model)
fp16_cfg = cfg.get('fp16', None)
if fp16_cfg is not None:
wrap_fp16_model(model)
load_checkpoint(model, args.checkpoint, map_location='cpu')
if args.fuse_conv_bn:
model = fuse_conv_bn(model)
if not distributed:
model = MMDataParallel(model, device_ids=[0])
is_kie = cfg.model.type in ['SDMGR']
outputs = single_gpu_test(model, data_loader, args.show, args.show_dir,
is_kie, args.show_score_thr)
else:
model = MMDistributedDataParallel(
model.cuda(),
device_ids=[torch.cuda.current_device()],
broadcast_buffers=False)
outputs = multi_gpu_test(model, data_loader, args.tmpdir,
args.gpu_collect)
rank, _ = get_dist_info()
if rank == 0:
if args.out:
print(f'\nwriting results to {args.out}')
mmcv.dump(outputs, args.out)
kwargs = {} if args.eval_options is None else args.eval_options
if args.format_only:
dataset.format_results(outputs, **kwargs)
if args.eval:
eval_kwargs = cfg.get('evaluation', {}).copy()
# hard-code way to remove EvalHook args
for key in [
'interval', 'tmpdir', 'start', 'gpu_collect', 'save_best',
'rule'
]:
eval_kwargs.pop(key, None)
eval_kwargs.update(dict(metric=args.eval, **kwargs))
print(dataset.evaluate(outputs, **eval_kwargs))
def test_drrg(cfg_file):
model = _get_detector_cfg(cfg_file)
model['pretrained'] = None
from mmocr.models import build_detector
detector = build_detector(model)
detector = revert_sync_batchnorm(detector)
input_shape = (1, 3, 224, 224)
num_kernels = 1
mm_inputs = _demo_mm_inputs(num_kernels, input_shape)
imgs = mm_inputs.pop('imgs')
img_metas = mm_inputs.pop('img_metas')
gt_text_mask = mm_inputs.pop('gt_text_mask')
gt_center_region_mask = mm_inputs.pop('gt_center_region_mask')
gt_mask = mm_inputs.pop('gt_mask')
gt_top_height_map = mm_inputs.pop('gt_radius_map')
gt_bot_height_map = gt_top_height_map.copy()
gt_sin_map = mm_inputs.pop('gt_sin_map')
gt_cos_map = mm_inputs.pop('gt_cos_map')
num_rois = 32
x = np.random.randint(4, 224, (num_rois, 1))
y = np.random.randint(4, 224, (num_rois, 1))
h = 4 * np.ones((num_rois, 1))
w = 4 * np.ones((num_rois, 1))
angle = (np.random.random_sample((num_rois, 1)) * 2 - 1) * np.pi / 2
cos, sin = np.cos(angle), np.sin(angle)
comp_labels = np.random.randint(1, 3, (num_rois, 1))
num_rois = num_rois * np.ones((num_rois, 1))
comp_attribs = np.hstack([num_rois, x, y, h, w, cos, sin, comp_labels])
gt_comp_attribs = np.expand_dims(comp_attribs.astype(np.float32), axis=0)
# Test forward train
losses = detector.forward(
imgs,
img_metas,
gt_text_mask=gt_text_mask,
gt_center_region_mask=gt_center_region_mask,
gt_mask=gt_mask,
gt_top_height_map=gt_top_height_map,
gt_bot_height_map=gt_bot_height_map,
gt_sin_map=gt_sin_map,
gt_cos_map=gt_cos_map,
gt_comp_attribs=gt_comp_attribs)
assert isinstance(losses, dict)
# Test forward test
model['bbox_head']['in_channels'] = 6
model['bbox_head']['text_region_thr'] = 0.8
model['bbox_head']['center_region_thr'] = 0.8
detector = build_detector(model)
maps = torch.zeros((1, 6, 224, 224), dtype=torch.float)
maps[:, 0:2, :, :] = -10.
maps[:, 0, 60:100, 50:170] = 10.
maps[:, 1, 75:85, 60:160] = 10.
maps[:, 2, 75:85, 60:160] = 0.
maps[:, 3, 75:85, 60:160] = 1.
maps[:, 4, 75:85, 60:160] = 10.
maps[:, 5, 75:85, 60:160] = 10.
with torch.no_grad():
full_pass_weight = torch.zeros((6, 6, 1, 1))
for i in range(6):
full_pass_weight[i, i, 0, 0] = 1
detector.bbox_head.out_conv.weight.data = full_pass_weight
detector.bbox_head.out_conv.bias.data.fill_(0.)
outs = detector.bbox_head.single_test(maps)
boundaries = detector.bbox_head.get_boundary(*outs, img_metas, True)
assert len(boundaries) == 1
# Test show result
results = {'boundary_result': [[0, 0, 1, 0, 1, 1, 0, 1, 0.9]]}
img = np.random.rand(5, 5)
detector.show_result(img, results)
def build_model(config_file):
device = 'cpu'
model = init_detector(config_file, checkpoint=None, device=device)
model = revert_sync_batchnorm(model)
return model
def build_model(cfg):
model = build_detector(cfg.model, test_cfg=cfg.get('test_cfg'))
model = revert_sync_batchnorm(model)
model = MMDataParallel(model)
return model
