def __init__(self,
low=None,
high=None,
seed=0,
dtype=mstype.float32,
name="Uniform"):
"""
Constructor of Uniform distribution.
"""
param = dict(locals())
super(Uniform, self).__init__(dtype, name, param)
if low is not None and high is not None:
self._low = convert_to_batch(low, self._broadcast_shape, dtype)
self._high = convert_to_batch(high, self._broadcast_shape, dtype)
check_greater(self.low, self.high, "low value", "high value")
else:
self._low = low
self._high = high
# ops needed for the class
self.const = P.ScalarToArray()
self.dtypeop = P.DType()
self.exp = P.Exp()
self.fill = P.Fill()
self.less = P.Less()
self.lessequal = P.LessEqual()
self.log = P.Log()
self.logicaland = P.LogicalAnd()
self.select = P.Select()
self.shape = P.Shape()
self.sq = P.Square()
self.sqrt = P.Sqrt()
self.uniform = P.UniformReal(seed=seed)
self.zeroslike = P.ZerosLike()
def __init__(self, batch_size=4):
super(DiceLoss, self).__init__()
self.threshold0 = Tensor(0.5, mstype.float32)
self.zero_float32 = Tensor(0.0, mstype.float32)
self.k = int(640 * 640)
self.negative_one_int32 = Tensor(-1, mstype.int32)
self.batch_size = batch_size
self.concat = P.Concat()
self.less_equal = P.LessEqual()
self.greater = P.Greater()
self.reduce_sum = P.ReduceSum()
self.reduce_sum_keep_dims = P.ReduceSum(keep_dims=True)
self.reduce_mean = P.ReduceMean()
self.reduce_min = P.ReduceMin()
self.cast = P.Cast()
self.minimum = P.Minimum()
self.expand_dims = P.ExpandDims()
self.select = P.Select()
self.fill = P.Fill()
self.topk = P.TopK(sorted=True)
self.shape = P.Shape()
self.sigmoid = P.Sigmoid()
self.reshape = P.Reshape()
self.slice = P.Slice()
self.logical_and = P.LogicalAnd()
self.logical_or = P.LogicalOr()
self.equal = P.Equal()
self.zeros_like = P.ZerosLike()
self.add = P.TensorAdd()
self.gather = P.Gather()
def __init__(self):
super(IGamma, self).__init__()
# const numbers
# If more data types are supported, this float max value need to be selected.
self.log_maxfloat32 = Tensor(np.log(np.finfo(np.float32).max),
mstype.float32)
# operations
self.logicaland = P.LogicalAnd()
self.logicalor = P.LogicalOr()
self.logicalnot = P.LogicalNot()
self.equal = P.Equal()
self.greater = P.Greater()
self.less = P.Less()
self.neg = P.Neg()
self.log = P.Log()
self.exp = P.Exp()
self.select = P.Select()
self.zeroslike = P.ZerosLike()
self.fill = P.Fill()
self.shape = P.Shape()
self.dtype = P.DType()
self.lgamma = LGamma()
self.const = P.ScalarToArray()
self.cast = P.Cast()
def __init__(self):
super(DiGamma, self).__init__()
# const numbers
self.k_lanczos_gamma = 7
self.k_base_lanczos_coeff = 0.99999999999980993227684700473478
self.k_lanczos_coefficients = [676.520368121885098567009190444019,
-1259.13921672240287047156078755283,
771.3234287776530788486528258894,
-176.61502916214059906584551354,
12.507343278686904814458936853,
-0.13857109526572011689554707,
9.984369578019570859563e-6,
1.50563273514931155834e-7]
self.nan = np.nan
self.pi = np.pi
self.lanczos_gamma_plus_one_half = self.k_lanczos_gamma + 0.5
self.log_lanczos_gamma_plus_one_half = np.log(self.lanczos_gamma_plus_one_half)
# operations
self.log1p = P.Log1p()
self.abs = P.Abs()
self.shape = P.Shape()
self.dtype = P.DType()
self.fill = P.Fill()
self.floor = P.Floor()
self.equal = P.Equal()
self.less = P.Less()
self.select = P.Select()
self.sin = P.Sin()
self.cos = P.Cos()
self.logicaland = P.LogicalAnd()
def __init__(self,
config,
batch_size,
num_classes,
target_means=(0., 0., 0., 0.),
target_stds=(0.1, 0.1, 0.2, 0.2)
):
super(RcnnMask, self).__init__()
cfg = config
self.rcnn_loss_mask_fb_weight = Tensor(np.array(cfg.rcnn_loss_mask_fb_weight).astype(np.float16))
self.rcnn_mask_out_channels = cfg.rcnn_mask_out_channels
self.target_means = target_means
self.target_stds = target_stds
self.num_classes = num_classes
self.in_channels = cfg.rcnn_in_channels
self.fpn_mask = FpnMask(self.in_channels, self.rcnn_mask_out_channels, self.num_classes)
self.logicaland = P.LogicalAnd()
self.loss_mask = P.SigmoidCrossEntropyWithLogits()
self.onehot = P.OneHot()
self.greater = P.Greater()
self.cast = P.Cast()
self.sum_loss = P.ReduceSum()
self.tile = P.Tile()
self.expandims = P.ExpandDims()
self.on_value = Tensor(1.0, mstype.float32)
self.off_value = Tensor(0.0, mstype.float32)
self.num_bboxes = cfg.num_expected_pos_stage2 * batch_size
rmv_first = np.ones((self.num_bboxes, self.num_classes))
rmv_first[:, 0] = np.zeros((self.num_bboxes,))
self.rmv_first_tensor = Tensor(rmv_first.astype(np.float16))
self.mean_loss = P.ReduceMean()
def __init__(self):
super(MathBinaryNet2, self).__init__()
self.less_equal = P.LessEqual()
self.greater = P.Greater()
self.logic_or = P.LogicalOr()
self.logic_and = P.LogicalAnd()
self.number = 3
self.flag = True
def __init__(self,
probs=None,
seed=None,
dtype=mstype.int32,
name="Categorical"):
param = dict(locals())
param['param_dict'] = {'probs': probs}
valid_dtype = mstype.uint_type + mstype.int_type + mstype.float_type
Validator.check_type_name("dtype", dtype, valid_dtype,
type(self).__name__)
super(Categorical, self).__init__(seed, dtype, name, param)
self._probs = self._add_parameter(probs, 'probs')
if self.probs is not None:
check_rank(self.probs)
check_prob(self.probs)
check_sum_equal_one(probs)
# update is_scalar_batch and broadcast_shape
# drop one dimension
if self.probs.shape[:-1] == ():
self._is_scalar_batch = True
self._broadcast_shape = self._broadcast_shape[:-1]
self.argmax = P.ArgMaxWithValue(axis=-1)
self.broadcast = broadcast_to
self.cast = P.Cast()
self.clip_by_value = C.clip_by_value
self.concat = P.Concat(-1)
self.cumsum = P.CumSum()
self.dtypeop = P.DType()
self.exp = exp_generic
self.expand_dim = P.ExpandDims()
self.fill = P.Fill()
self.gather = P.GatherNd()
self.greater = P.Greater()
self.issubclass = P.IsSubClass()
self.less = P.Less()
self.log = log_generic
self.log_softmax = P.LogSoftmax()
self.logicor = P.LogicalOr()
self.logicand = P.LogicalAnd()
self.multinomial = P.Multinomial(seed=self.seed)
self.reshape = P.Reshape()
self.reduce_sum = P.ReduceSum(keep_dims=True)
self.select = P.Select()
self.shape = P.Shape()
self.softmax = P.Softmax()
self.squeeze = P.Squeeze()
self.squeeze_first_axis = P.Squeeze(0)
self.squeeze_last_axis = P.Squeeze(-1)
self.square = P.Square()
self.transpose = P.Transpose()
self.is_nan = P.IsNan()
self.index_type = mstype.int32
self.nan = np.nan
def test_logicaland():
op = P.LogicalAnd()
op_wrapper = OpNetWrapper(op)
input_x = Tensor(np.array([True, False, False]))
input_y = Tensor(np.array([True, True, False]))
outputs = op_wrapper(input_x, input_y)
assert np.allclose(outputs.asnumpy(), (True, False, False))
def __init__(self, config, batch_size, num_bboxes, add_gt_as_proposals):
super(BboxAssignSample, self).__init__()
cfg = config
self.batch_size = batch_size
self.neg_iou_thr = Tensor(cfg.neg_iou_thr, mstype.float16)
self.pos_iou_thr = Tensor(cfg.pos_iou_thr, mstype.float16)
self.min_pos_iou = Tensor(cfg.min_pos_iou, mstype.float16)
self.zero_thr = Tensor(0.0, mstype.float16)
self.num_bboxes = num_bboxes
self.num_gts = cfg.num_gts
self.num_expected_pos = cfg.num_expected_pos
self.num_expected_neg = cfg.num_expected_neg
self.add_gt_as_proposals = add_gt_as_proposals
if self.add_gt_as_proposals:
self.label_inds = Tensor(np.arange(1, self.num_gts + 1))
self.concat = P.Concat(axis=0)
self.max_gt = P.ArgMaxWithValue(axis=0)
self.max_anchor = P.ArgMaxWithValue(axis=1)
self.sum_inds = P.ReduceSum()
self.iou = P.IOU()
self.greaterequal = P.GreaterEqual()
self.greater = P.Greater()
self.select = P.Select()
self.gatherND = P.GatherNd()
self.squeeze = P.Squeeze()
self.cast = P.Cast()
self.logicaland = P.LogicalAnd()
self.less = P.Less()
self.random_choice_with_mask_pos = P.RandomChoiceWithMask(self.num_expected_pos)
self.random_choice_with_mask_neg = P.RandomChoiceWithMask(self.num_expected_neg)
self.reshape = P.Reshape()
self.equal = P.Equal()
self.bounding_box_encode = BoundingBoxEncode()
self.scatterNdUpdate = P.ScatterNdUpdate()
self.scatterNd = P.ScatterNd()
self.logicalnot = P.LogicalNot()
self.tile = P.Tile()
self.zeros_like = P.ZerosLike()
self.assigned_gt_inds = Tensor(np.array(-1 * np.ones(num_bboxes), dtype=np.int32))
self.assigned_gt_zeros = Tensor(np.array(np.zeros(num_bboxes), dtype=np.int32))
self.assigned_gt_ones = Tensor(np.array(np.ones(num_bboxes), dtype=np.int32))
self.assigned_gt_ignores = Tensor(np.array(-1 * np.ones(num_bboxes), dtype=np.int32))
self.assigned_pos_ones = Tensor(np.array(np.ones(self.num_expected_pos), dtype=np.int32))
self.check_neg_mask = Tensor(np.array(np.ones(self.num_expected_neg - self.num_expected_pos), dtype=np.bool))
self.range_pos_size = Tensor(np.arange(self.num_expected_pos).astype(np.float16))
self.check_gt_one = Tensor(np.array(-1 * np.ones((self.num_gts, 4)), dtype=np.float16))
self.check_anchor_two = Tensor(np.array(-2 * np.ones((self.num_bboxes, 4)), dtype=np.float16))
self.print = P.Print()
def __init__(self,
concentration1=None,
concentration0=None,
seed=None,
dtype=mstype.float32,
name="Beta"):
"""
Constructor of Beta.
"""
param = dict(locals())
param['param_dict'] = {
'concentration1': concentration1,
'concentration0': concentration0
}
valid_dtype = mstype.float_type
Validator.check_type_name("dtype", dtype, valid_dtype,
type(self).__name__)
# As some operators can't accept scalar input, check the type here
if isinstance(concentration0, float):
raise TypeError("Input concentration0 can't be scalar")
if isinstance(concentration1, float):
raise TypeError("Input concentration1 can't be scalar")
super(Beta, self).__init__(seed, dtype, name, param)
self._concentration1 = self._add_parameter(concentration1,
'concentration1')
self._concentration0 = self._add_parameter(concentration0,
'concentration0')
if self._concentration1 is not None:
check_greater_zero(self._concentration1, "concentration1")
if self._concentration0 is not None:
check_greater_zero(self._concentration0, "concentration0")
# ops needed for the class
self.log = log_generic
self.log1p = P.Log1p()
self.neg = P.Neg()
self.pow = P.Pow()
self.squeeze = P.Squeeze(0)
self.cast = P.Cast()
self.fill = P.Fill()
self.shape = P.Shape()
self.select = P.Select()
self.logicaland = P.LogicalAnd()
self.greater = P.Greater()
self.digamma = nn.DiGamma()
self.lbeta = nn.LBeta()
def __init__(self,
config,
batch_size,
num_classes,
target_means=(0., 0., 0., 0.),
target_stds=(0.1, 0.1, 0.2, 0.2)):
super(RcnnCls, self).__init__()
cfg = config
self.rcnn_loss_cls_weight = Tensor(
np.array(cfg.rcnn_loss_cls_weight).astype(np.float16))
self.rcnn_loss_reg_weight = Tensor(
np.array(cfg.rcnn_loss_reg_weight).astype(np.float16))
self.rcnn_fc_out_channels = cfg.rcnn_fc_out_channels
self.target_means = target_means
self.target_stds = target_stds
self.num_classes = num_classes
self.in_channels = cfg.rcnn_in_channels
self.train_batch_size = batch_size
self.test_batch_size = cfg.test_batch_size
self.fpn_cls = FpnCls(self.in_channels, self.rcnn_fc_out_channels,
self.num_classes, cfg.roi_layer["out_size"])
self.relu = P.ReLU()
self.logicaland = P.LogicalAnd()
self.loss_cls = P.SoftmaxCrossEntropyWithLogits()
self.loss_bbox = P.SmoothL1Loss(beta=1.0)
self.loss_mask = P.SigmoidCrossEntropyWithLogits()
self.reshape = P.Reshape()
self.onehot = P.OneHot()
self.greater = P.Greater()
self.cast = P.Cast()
self.sum_loss = P.ReduceSum()
self.tile = P.Tile()
self.expandims = P.ExpandDims()
self.gather = P.GatherNd()
self.argmax = P.ArgMaxWithValue(axis=1)
self.on_value = Tensor(1.0, mstype.float32)
self.off_value = Tensor(0.0, mstype.float32)
self.value = Tensor(1.0, mstype.float16)
self.num_bboxes = (cfg.num_expected_pos_stage2 +
cfg.num_expected_neg_stage2) * batch_size
rmv_first = np.ones((self.num_bboxes, self.num_classes))
rmv_first[:, 0] = np.zeros((self.num_bboxes, ))
self.rmv_first_tensor = Tensor(rmv_first.astype(np.float16))
self.num_bboxes_test = cfg.rpn_max_num * cfg.test_batch_size
def __init__(self,
low=None,
high=None,
seed=None,
dtype=mstype.float32,
name="Uniform"):
"""
Constructor of Uniform distribution.
"""
param = dict(locals())
valid_dtype = mstype.float_type
check_type(dtype, valid_dtype, type(self).__name__)
super(Uniform, self).__init__(seed, dtype, name, param)
self.parameter_type = set_param_type({
'low': low,
'high': high
}, self.dtype)
if low is not None and high is not None:
self._low = cast_to_tensor(low, self.parameter_type)
self._high = cast_to_tensor(high, self.parameter_type)
check_greater(self.low, self.high, "low value", "high value")
else:
self._low = low if low is None else cast_to_tensor(
low, self.parameter_type)
self._high = high if high is None else cast_to_tensor(
high, self.parameter_type)
self.default_parameters = [self.low, self.high]
self.parameter_names = ['low', 'high']
# ops needed for the class
self.exp = exp_generic
self.log = log_generic
self.squeeze = P.Squeeze(0)
self.cast = P.Cast()
self.const = P.ScalarToArray()
self.dtypeop = P.DType()
self.fill = P.Fill()
self.less = P.Less()
self.lessequal = P.LessEqual()
self.logicaland = P.LogicalAnd()
self.select = P.Select()
self.shape = P.Shape()
self.sq = P.Square()
self.sqrt = P.Sqrt()
self.zeroslike = P.ZerosLike()
self.uniform = C.uniform
self.sametypeshape = P.SameTypeShape()
def _IgammaSeries(ax, x, a, enabled):
"""Helper function for computing Igamma using a power series."""
logicaland = P.LogicalAnd()
greater = P.Greater()
fill = P.Fill()
shape = P.Shape()
dtype = P.DType()
select = P.Select()
if dtype(ax) == mstype.float16:
epsilon = eps_fp16
else:
epsilon = eps_fp32
def cond(vals):
enabled = vals[0]
return enabled
def body(vals):
enabled = vals[0]
r = vals[1]
c = vals[2]
ans = vals[3]
x = vals[4]
dc_da = vals[5]
dans_da = vals[6]
r = r + 1
dc_da = dc_da * (x / r) + (-1 * c * x) / (r * r)
dans_da = dans_da + dc_da
c = c * (x / r)
ans = ans + c
conditional = logicaland(enabled, greater(c / ans, epsilon))
return (conditional, select(enabled, r,
vals[1]), select(enabled, c, vals[2]),
select(enabled, ans, vals[3]), select(enabled, x, vals[4]),
select(enabled, dc_da,
vals[5]), select(enabled, dans_da, vals[6]))
ones = fill(dtype(a), shape(a), 1)
zeros = fill(dtype(a), shape(a), 0)
vals = (enabled, a, ones, ones, x, zeros, zeros)
vals = _while_helper_func(cond, body, vals)
ans = vals[3]
return (ans * ax) / a
def __init__(self, config):
super(ETSNet, self).__init__()
self.kernel_num = config.KERNEL_NUM
self.inference = config.INFERENCE
if config.INFERENCE:
self.long_size = config.INFER_LONG_SIZE
else:
self.long_size = config.TRAIN_LONG_SIZE
# backbone
self.feature_extractor = ResNet(ResidualBlock,
config.BACKBONE_LAYER_NUMS,
config.BACKBONE_IN_CHANNELS,
config.BACKBONE_OUT_CHANNELS)
# neck
self.feature_fusion = FPN(config.BACKBONE_OUT_CHANNELS,
config.NECK_OUT_CHANNEL, self.long_size)
# head
self.conv1 = _conv(4 * config.NECK_OUT_CHANNEL,
config.NECK_OUT_CHANNEL,
kernel_size=3,
stride=1,
has_bias=True)
self.bn1 = _bn(config.NECK_OUT_CHANNEL)
self.relu1 = nn.ReLU()
self.conv2 = _conv(config.NECK_OUT_CHANNEL,
config.KERNEL_NUM,
kernel_size=1,
has_bias=True)
self._upsample = P.ResizeBilinear((self.long_size, self.long_size),
align_corners=True)
if self.inference:
self.one_float32 = Tensor(1.0, mstype.float32)
self.sigmoid = P.Sigmoid()
self.greater = P.Greater()
self.logic_and = P.LogicalAnd()
print('ETSNet initialized!')
def test_mode_init(self, config):
self.test_batch_size = config.test_batch_size
self.split = P.Split(axis=0, output_num=self.test_batch_size)
self.split_shape = P.Split(axis=0, output_num=4)
self.split_scores = P.Split(axis=1, output_num=self.num_classes)
self.split_cls = P.Split(axis=0, output_num=self.num_classes - 1)
self.tile = P.Tile()
self.gather = P.GatherNd()
self.rpn_max_num = config.rpn_max_num
self.zeros_for_nms = Tensor(
np.zeros((self.rpn_max_num, 3)).astype(self.dtype))
self.ones_mask = np.ones((self.rpn_max_num, 1)).astype(np.bool)
self.zeros_mask = np.zeros((self.rpn_max_num, 1)).astype(np.bool)
self.bbox_mask = Tensor(
np.concatenate((self.ones_mask, self.zeros_mask, self.ones_mask,
self.zeros_mask),
axis=1))
self.nms_pad_mask = Tensor(
np.concatenate((self.ones_mask, self.ones_mask, self.ones_mask,
self.ones_mask, self.zeros_mask),
axis=1))
self.test_score_thresh = Tensor(
np.ones((self.rpn_max_num, 1)).astype(self.dtype) *
config.test_score_thr)
self.test_score_zeros = Tensor(
np.ones((self.rpn_max_num, 1)).astype(self.dtype) * 0)
self.test_box_zeros = Tensor(
np.ones((self.rpn_max_num, 4)).astype(self.dtype) * -1)
self.test_iou_thr = Tensor(
np.ones((self.rpn_max_num, 1)).astype(self.dtype) *
config.test_iou_thr)
self.test_max_per_img = config.test_max_per_img
self.nms_test = P.NMSWithMask(config.test_iou_thr)
self.softmax = P.Softmax(axis=1)
self.logicand = P.LogicalAnd()
self.oneslike = P.OnesLike()
self.test_topk = P.TopK(sorted=True)
self.test_num_proposal = self.test_batch_size * self.rpn_max_num
def __init__(self,
low=None,
high=None,
seed=None,
dtype=mstype.float32,
name="Uniform"):
"""
Constructor of Uniform distribution.
"""
param = dict(locals())
param['param_dict'] = {'low': low, 'high': high}
valid_dtype = mstype.float_type
Validator.check_type_name("dtype", dtype, valid_dtype,
type(self).__name__)
super(Uniform, self).__init__(seed, dtype, name, param)
self._low = self._add_parameter(low, 'low')
self._high = self._add_parameter(high, 'high')
if self.low is not None and self.high is not None:
check_greater(self.low, self.high, 'low', 'high')
# ops needed for the class
self.exp = exp_generic
self.log = log_generic
self.squeeze = P.Squeeze(0)
self.cast = P.Cast()
self.const = P.ScalarToArray()
self.dtypeop = P.DType()
self.fill = P.Fill()
self.less = P.Less()
self.lessequal = P.LessEqual()
self.logicaland = P.LogicalAnd()
self.select = P.Select()
self.shape = P.Shape()
self.sq = P.Square()
self.zeroslike = P.ZerosLike()
self.uniform = C.uniform
def __init__(self,
config,
representation_size,
batch_size,
num_classes,
target_means=(0., 0., 0., 0.),
target_stds=(0.1, 0.1, 0.2, 0.2)):
super(Rcnn, self).__init__()
cfg = config
self.rcnn_loss_cls_weight = Tensor(
np.array(cfg.rcnn_loss_cls_weight).astype(np.float16))
self.rcnn_loss_reg_weight = Tensor(
np.array(cfg.rcnn_loss_reg_weight).astype(np.float16))
self.rcnn_fc_out_channels = cfg.rcnn_fc_out_channels
self.target_means = target_means
self.target_stds = target_stds
self.num_classes = num_classes
self.in_channels = cfg.rcnn_in_channels
self.train_batch_size = batch_size
self.test_batch_size = cfg.test_batch_size
self.use_ambigous_sample = cfg.use_ambigous_sample
shape_0 = (self.rcnn_fc_out_channels, representation_size)
weights_0 = initializer("XavierUniform",
shape=shape_0[::-1],
dtype=mstype.float16).to_tensor()
shape_1 = (self.rcnn_fc_out_channels, self.rcnn_fc_out_channels)
weights_1 = initializer("XavierUniform",
shape=shape_1[::-1],
dtype=mstype.float16).to_tensor()
self.shared_fc_0 = DenseNoTranpose(representation_size,
self.rcnn_fc_out_channels,
weights_0)
self.shared_fc_1 = DenseNoTranpose(self.rcnn_fc_out_channels,
self.rcnn_fc_out_channels,
weights_1)
cls_weight = initializer(
'Normal',
shape=[num_classes, self.rcnn_fc_out_channels][::-1],
dtype=mstype.float16).to_tensor()
reg_weight = initializer(
'Normal',
shape=[num_classes * 4, self.rcnn_fc_out_channels][::-1],
dtype=mstype.float16).to_tensor()
self.cls_scores = DenseNoTranpose(self.rcnn_fc_out_channels,
num_classes, cls_weight)
self.reg_scores = DenseNoTranpose(self.rcnn_fc_out_channels,
num_classes * 4, reg_weight)
self.flatten = P.Flatten()
self.relu = P.ReLU()
self.logicaland = P.LogicalAnd()
self.loss_cls = P.SoftmaxCrossEntropyWithLogits()
self.loss_bbox = P.SmoothL1Loss(beta=1.0)
self.reshape = P.Reshape()
self.onehot = P.OneHot()
self.greater = P.Greater()
self.equal = P.Equal()
self.cast = P.Cast()
self.sum_loss = P.ReduceSum()
self.tile = P.Tile()
self.expandims = P.ExpandDims()
self.gather = P.GatherNd()
self.argmax = P.ArgMaxWithValue(axis=1)
self.on_value = Tensor(1.0, mstype.float32)
self.off_value = Tensor(0.0, mstype.float32)
self.value = Tensor(1.0, mstype.float16)
self.num_bboxes = (cfg.num_expected_pos_stage2 +
cfg.num_expected_neg_stage2) * batch_size
if self.use_ambigous_sample:
self.num_bboxes = (cfg.num_expected_pos_stage2 +
cfg.num_expected_amb_stage2 +
cfg.num_expected_neg_stage2) * batch_size
rmv_first = np.ones((self.num_bboxes, self.num_classes))
rmv_first[:, 0] = np.zeros((self.num_bboxes, ))
self.rmv_first_tensor = Tensor(rmv_first.astype(np.float16))
self.num_bboxes_test = cfg.rpn_max_num * cfg.test_batch_size
range_max = np.arange(self.num_bboxes_test).astype(np.int32)
self.range_max = Tensor(range_max)
def __init__(self):
super(NetAnd, self).__init__()
self.logicaland = P.LogicalAnd()
def __init__(self, config, batch_size, num_bboxes, add_gt_as_proposals):
super(BboxAssignSampleForRcnn, self).__init__()
cfg = config
self.batch_size = batch_size
self.neg_iou_thr = cfg.neg_iou_thr_stage2
self.pos_iou_thr = cfg.pos_iou_thr_stage2
self.min_pos_iou = cfg.min_pos_iou_stage2
self.num_gts = cfg.num_gts
self.num_bboxes = num_bboxes
self.num_expected_pos = cfg.num_expected_pos_stage2
self.num_expected_neg = cfg.num_expected_neg_stage2
self.num_expected_total = cfg.num_expected_total_stage2
self.add_gt_as_proposals = add_gt_as_proposals
self.label_inds = Tensor(
np.arange(1, self.num_gts + 1).astype(np.int32))
self.add_gt_as_proposals_valid = Tensor(
np.array(self.add_gt_as_proposals * np.ones(self.num_gts),
dtype=np.int32))
self.concat = P.Concat(axis=0)
self.max_gt = P.ArgMaxWithValue(axis=0)
self.max_anchor = P.ArgMaxWithValue(axis=1)
self.sum_inds = P.ReduceSum()
self.iou = P.IOU()
self.greaterequal = P.GreaterEqual()
self.greater = P.Greater()
self.select = P.Select()
self.gatherND = P.GatherNd()
self.squeeze = P.Squeeze()
self.cast = P.Cast()
self.logicaland = P.LogicalAnd()
self.less = P.Less()
self.random_choice_with_mask_pos = P.RandomChoiceWithMask(
self.num_expected_pos)
self.random_choice_with_mask_neg = P.RandomChoiceWithMask(
self.num_expected_neg)
self.reshape = P.Reshape()
self.equal = P.Equal()
self.bounding_box_encode = P.BoundingBoxEncode(means=(0.0, 0.0, 0.0,
0.0),
stds=(0.1, 0.1, 0.2,
0.2))
self.concat_axis1 = P.Concat(axis=1)
self.logicalnot = P.LogicalNot()
self.tile = P.Tile()
# Check
self.check_gt_one = Tensor(
np.array(-1 * np.ones((self.num_gts, 4)), dtype=np.float16))
self.check_anchor_two = Tensor(
np.array(-2 * np.ones((self.num_bboxes, 4)), dtype=np.float16))
# Init tensor
self.assigned_gt_inds = Tensor(
np.array(-1 * np.ones(num_bboxes), dtype=np.int32))
self.assigned_gt_zeros = Tensor(
np.array(np.zeros(num_bboxes), dtype=np.int32))
self.assigned_gt_ones = Tensor(
np.array(np.ones(num_bboxes), dtype=np.int32))
self.assigned_gt_ignores = Tensor(
np.array(-1 * np.ones(num_bboxes), dtype=np.int32))
self.assigned_pos_ones = Tensor(
np.array(np.ones(self.num_expected_pos), dtype=np.int32))
self.gt_ignores = Tensor(
np.array(-1 * np.ones(self.num_gts), dtype=np.int32))
self.range_pos_size = Tensor(
np.arange(self.num_expected_pos).astype(np.float16))
self.check_neg_mask = Tensor(
np.array(np.ones(self.num_expected_neg - self.num_expected_pos),
dtype=np.bool))
self.bboxs_neg_mask = Tensor(
np.zeros((self.num_expected_neg, 4), dtype=np.float16))
self.labels_neg_mask = Tensor(
np.array(np.zeros(self.num_expected_neg), dtype=np.uint8))
self.reshape_shape_pos = (self.num_expected_pos, 1)
self.reshape_shape_neg = (self.num_expected_neg, 1)
self.scalar_zero = Tensor(0.0, dtype=mstype.float16)
self.scalar_neg_iou_thr = Tensor(self.neg_iou_thr,
dtype=mstype.float16)
self.scalar_pos_iou_thr = Tensor(self.pos_iou_thr,
dtype=mstype.float16)
self.scalar_min_pos_iou = Tensor(self.min_pos_iou,
dtype=mstype.float16)
self.expand_dims = P.ExpandDims()
self.split = P.Split(axis=1, output_num=4)
self.concat_last_axis = P.Concat(axis=-1)
self.round = P.Round()
self.image_h_w = Tensor(
[cfg.img_height, cfg.img_width, cfg.img_height, cfg.img_width],
dtype=mstype.float16)
self.range = nn.Range(start=0, limit=cfg.num_expected_pos_stage2)
self.crop_and_resize = P.CropAndResize(method="bilinear_v2")
self.mask_shape = (cfg.mask_shape[0], cfg.mask_shape[1])
self.squeeze_mask_last = P.Squeeze(axis=-1)
'desc_bprop': [Tensor(np.ones((2, 3, 4, 5), np.bool_))],
'skip': ['backward']}),
('Greater', {
'block': P.Greater(),
'desc_inputs': [[2, 3, 4, 1], [4, 5]],
'desc_bprop': [Tensor(np.ones((2, 3, 4, 5), np.bool_))]}),
('GreaterEqual', {
'block': P.GreaterEqual(),
'desc_inputs': [[2, 3, 4, 1], [4, 5]],
'desc_bprop': [Tensor(np.ones((2, 3, 4, 5), np.bool_))]}),
('LogicalNot', {
'block': P.LogicalNot(),
'desc_inputs': [Tensor(np.zeros((3, 4, 5), np.bool_))],
'desc_bprop': [Tensor(np.ones((3, 4, 5), np.bool_))]}),
('LogicalAnd', {
'block': P.LogicalAnd(),
'desc_inputs': [Tensor(np.zeros((2, 3, 4), np.bool_)), Tensor(np.ones((1), np.bool_))],
'desc_bprop': [Tensor(np.zeros((2, 3, 4), np.bool_))]}),
('LogicalOr', {
'block': P.LogicalOr(),
'desc_inputs': [Tensor(np.zeros((3, 4, 5), np.bool_)), Tensor(np.ones((3, 1, 1), np.bool_))],
'desc_bprop': [Tensor(np.zeros((3, 4, 5), np.bool_))]}),
('NpuAllocFloatStatus', {
'block': P.NPUAllocFloatStatus(),
'desc_inputs': [],
'add_fack_input': True,
'fack_input_type': np.float32,
'desc_bprop': [Tensor(np.zeros([8]).astype(np.float32))],
'skip': ['backward']}),
('NpuGetFloatStatus', {
'block': P.NPUGetFloatStatus(),
'desc_inputs': [[2, 3, 4, 1], [4, 5]],
'desc_bprop': [Tensor(np.ones((2, 3, 4, 5), np.bool_))]
}),
('GreaterEqual', {
'block': P.GreaterEqual(),
'desc_inputs': [[2, 3, 4, 1], [4, 5]],
'desc_bprop': [Tensor(np.ones((2, 3, 4, 5), np.bool_))]
}),
('LogicalNot', {
'block': P.LogicalNot(),
'desc_inputs': [Tensor(np.zeros((3, 4, 5), np.bool_))],
'desc_bprop': [Tensor(np.ones((3, 4, 5), np.bool_))]
}),
('LogicalAnd', {
'block':
P.LogicalAnd(),
'desc_inputs': [
Tensor(np.zeros((2, 3, 4), np.bool_)),
Tensor(np.ones((1), np.bool_))
],
'desc_bprop': [Tensor(np.zeros((2, 3, 4), np.bool_))]
}),
('LogicalOr', {
'block':
P.LogicalOr(),
'desc_inputs': [
Tensor(np.zeros((3, 4, 5), np.bool_)),
Tensor(np.ones((3, 1, 1), np.bool_))
],
'desc_bprop': [Tensor(np.zeros((3, 4, 5), np.bool_))]
}),
def __init__(self, config):
super(Deeptext_VGG16, self).__init__()
self.train_batch_size = config.batch_size
self.num_classes = config.num_classes
self.anchor_scales = config.anchor_scales
self.anchor_ratios = config.anchor_ratios
self.anchor_strides = config.anchor_strides
self.target_means = tuple(config.rcnn_target_means)
self.target_stds = tuple(config.rcnn_target_stds)
# Anchor generator
anchor_base_sizes = None
self.anchor_base_sizes = list(
self.anchor_strides
) if anchor_base_sizes is None else anchor_base_sizes
self.anchor_generators = []
for anchor_base in self.anchor_base_sizes:
self.anchor_generators.append(
AnchorGenerator(anchor_base, self.anchor_scales,
self.anchor_ratios))
self.num_anchors = len(self.anchor_ratios) * len(self.anchor_scales)
featmap_sizes = config.feature_shapes
assert len(featmap_sizes) == len(self.anchor_generators)
self.anchor_list = self.get_anchors(featmap_sizes)
# Rpn and rpn loss
self.gt_labels_stage1 = Tensor(
np.ones((self.train_batch_size, config.num_gts)).astype(np.uint8))
self.rpn_with_loss = RPN(config, self.train_batch_size,
config.rpn_in_channels,
config.rpn_feat_channels, config.num_anchors,
config.rpn_cls_out_channels)
# Proposal
self.proposal_generator = Proposal(config, self.train_batch_size,
config.activate_num_classes,
config.use_sigmoid_cls)
self.proposal_generator.set_train_local(config, True)
self.proposal_generator_test = Proposal(config, config.test_batch_size,
config.activate_num_classes,
config.use_sigmoid_cls)
self.proposal_generator_test.set_train_local(config, False)
# Assign and sampler stage two
self.bbox_assigner_sampler_for_rcnn = BboxAssignSampleForRcnn(
config, self.train_batch_size, config.num_bboxes_stage2, True)
self.decode = P.BoundingBoxDecode(max_shape=(576, 960), means=self.target_means, \
stds=self.target_stds)
# Rcnn
self.rcnn = Rcnn(
config, config.rcnn_in_channels * config.roi_layer['out_size'] *
config.roi_layer['out_size'], self.train_batch_size,
self.num_classes)
# Op declare
self.squeeze = P.Squeeze()
self.cast = P.Cast()
self.concat = P.Concat(axis=0)
self.concat_1 = P.Concat(axis=1)
self.concat_2 = P.Concat(axis=2)
self.reshape = P.Reshape()
self.select = P.Select()
self.greater = P.Greater()
self.transpose = P.Transpose()
# Test mode
self.test_batch_size = config.test_batch_size
self.split = P.Split(axis=0, output_num=self.test_batch_size)
self.split_shape = P.Split(axis=0, output_num=4)
self.split_scores = P.Split(axis=1, output_num=self.num_classes)
self.split_cls = P.Split(axis=0, output_num=self.num_classes - 1)
self.tile = P.Tile()
self.gather = P.GatherNd()
self.rpn_max_num = config.rpn_max_num
self.zeros_for_nms = Tensor(
np.zeros((self.rpn_max_num, 3)).astype(np.float32))
self.ones_mask = np.ones((self.rpn_max_num, 1)).astype(np.bool)
self.zeros_mask = np.zeros((self.rpn_max_num, 1)).astype(np.bool)
self.bbox_mask = Tensor(
np.concatenate((self.ones_mask, self.zeros_mask, self.ones_mask,
self.zeros_mask),
axis=1))
self.nms_pad_mask = Tensor(
np.concatenate((self.ones_mask, self.ones_mask, self.ones_mask,
self.ones_mask, self.zeros_mask),
axis=1))
self.test_score_thresh = Tensor(
np.ones((self.rpn_max_num, 1)).astype(np.float32) *
config.test_score_thr)
self.test_score_zeros = Tensor(
np.ones((self.rpn_max_num, 1)).astype(np.float32) * 0)
self.test_box_zeros = Tensor(
np.ones((self.rpn_max_num, 4)).astype(np.float32) * -1)
self.test_iou_thr = Tensor(
np.ones((self.rpn_max_num, 1)).astype(np.float32) *
config.test_iou_thr)
self.test_max_per_img = config.test_max_per_img
self.nms_test = P.NMSWithMask(config.test_iou_thr)
self.softmax = P.Softmax(axis=1)
self.logicand = P.LogicalAnd()
self.oneslike = P.OnesLike()
self.test_topk = P.TopK(sorted=True)
self.test_num_proposal = self.test_batch_size * self.rpn_max_num
# Improve speed
self.concat_start = (self.num_classes - 2)
self.concat_end = (self.num_classes - 1)
# Init tensor
self.use_ambigous_sample = config.use_ambigous_sample
roi_align_index = [
np.array(np.ones((config.num_expected_pos_stage2 +
config.num_expected_neg_stage2, 1)) * i,
dtype=np.float32) for i in range(self.train_batch_size)
]
if self.use_ambigous_sample:
roi_align_index = [
np.array(np.ones((config.num_expected_pos_stage2 +
config.num_expected_amb_stage2 +
config.num_expected_neg_stage2, 1)) * i,
dtype=np.float32)
for i in range(self.train_batch_size)
]
roi_align_index_test = [np.array(np.ones((config.rpn_max_num, 1)) * i, dtype=np.float32) \
for i in range(self.test_batch_size)]
self.roi_align_index_tensor = Tensor(np.concatenate(roi_align_index))
self.roi_align_index_test_tensor = Tensor(
np.concatenate(roi_align_index_test))
self.roi_align4 = P.ROIAlign(pooled_width=7,
pooled_height=7,
spatial_scale=0.125)
self.roi_align5 = P.ROIAlign(pooled_width=7,
pooled_height=7,
spatial_scale=0.0625)
self.concat1 = P.Concat(axis=1)
self.roi_align_fuse = _conv(in_channels=1024,
out_channels=512,
kernel_size=1,
padding=0,
stride=1)
self.vgg16_feature_extractor = VGG16FeatureExtraction()
def __init__(self):
super(NetConditionLackBranch, self).__init__()
self.logicaland = P.LogicalAnd()
self.logicalor = P.LogicalOr()
from mindspore.nn import Optimizer
from mindspore.nn import TrainOneStepCell, WithLossCell
from mindspore.nn.optim import Momentum
from mindspore.train import Model
from ....dataset_mock import MindData
context.set_context(mode=context.GRAPH_MODE, enable_sparse=True)
reduce_sum = P.ReduceSum()
unsorted_segment_sum = P.UnsortedSegmentSum()
transpose = P.Transpose()
shape_op = P.Shape()
reshape = P.Reshape()
size_op = P.Size()
invert_permutation = P.InvertPermutation()
logical_and = P.LogicalAnd()
def get_axis(x):
shape = shape_op(x)
length = F.tuple_len(shape)
perm = F.make_range(0, length)
return perm
class MSELoss(nn.Cell):
def __init__(self):
super(MSELoss, self).__init__()
self.reduce_sum = P.ReduceSum()
self.square = P.Square()
self.reduce_mean = P.ReduceMean()
'skip': ['backward']
}),
# input x is not Tensor(bool)
('LogicalNot1', {
'block': (P.LogicalNot(), {
'exception': TypeError,
'error_keywords': ['LogicalNot']
}),
'desc_inputs': [Tensor(np.ones([2, 3]).astype(np.int32))],
'skip': ['backward']
}),
# type of x and y not match
('LogicalAnd1', {
'block': (P.LogicalAnd(), {
'exception': TypeError,
'error_keywords': ['LogicalAnd']
}),
'desc_inputs': [
Tensor(np.ones([3, 4]).astype(np.int32)),
Tensor(np.ones([3, 4]).astype(np.bool_))
],
'skip': ['backward']
}),
# shape of x and y not match
('LogicalAnd2', {
'block': (P.LogicalAnd(), {
'exception': ValueError,
'error_keywords': ['LogicalAnd']
}),
def __init__(self, config):
super(Mask_Rcnn_Resnet50, self).__init__()
self.train_batch_size = config.batch_size
self.num_classes = config.num_classes
self.anchor_scales = config.anchor_scales
self.anchor_ratios = config.anchor_ratios
self.anchor_strides = config.anchor_strides
self.target_means = tuple(config.rcnn_target_means)
self.target_stds = tuple(config.rcnn_target_stds)
# Anchor generator
anchor_base_sizes = None
self.anchor_base_sizes = list(
self.anchor_strides) if anchor_base_sizes is None else anchor_base_sizes
self.anchor_generators = []
for anchor_base in self.anchor_base_sizes:
self.anchor_generators.append(
AnchorGenerator(anchor_base, self.anchor_scales, self.anchor_ratios))
self.num_anchors = len(self.anchor_ratios) * len(self.anchor_scales)
featmap_sizes = config.feature_shapes
assert len(featmap_sizes) == len(self.anchor_generators)
self.anchor_list = self.get_anchors(featmap_sizes)
# Backbone resnet50
self.backbone = ResNetFea(ResidualBlockUsing,
config.resnet_block,
config.resnet_in_channels,
config.resnet_out_channels,
False)
# Fpn
self.fpn_ncek = FeatPyramidNeck(config.fpn_in_channels,
config.fpn_out_channels,
config.fpn_num_outs)
# Rpn and rpn loss
self.gt_labels_stage1 = Tensor(np.ones((self.train_batch_size, config.num_gts)).astype(np.uint8))
self.rpn_with_loss = RPN(config,
self.train_batch_size,
config.rpn_in_channels,
config.rpn_feat_channels,
config.num_anchors,
config.rpn_cls_out_channels)
# Proposal
self.proposal_generator = Proposal(config,
self.train_batch_size,
config.activate_num_classes,
config.use_sigmoid_cls)
self.proposal_generator.set_train_local(config, True)
self.proposal_generator_test = Proposal(config,
config.test_batch_size,
config.activate_num_classes,
config.use_sigmoid_cls)
self.proposal_generator_test.set_train_local(config, False)
# Assign and sampler stage two
self.bbox_assigner_sampler_for_rcnn = BboxAssignSampleForRcnn(config, self.train_batch_size,
config.num_bboxes_stage2, True)
self.decode = P.BoundingBoxDecode(max_shape=(768, 1280), means=self.target_means, \
stds=self.target_stds)
# Roi
self.roi_align = SingleRoIExtractor(config,
config.roi_layer,
config.roi_align_out_channels,
config.roi_align_featmap_strides,
self.train_batch_size,
config.roi_align_finest_scale,
mask=False)
self.roi_align.set_train_local(config, True)
self.roi_align_mask = SingleRoIExtractor(config,
config.roi_layer,
config.roi_align_out_channels,
config.roi_align_featmap_strides,
self.train_batch_size,
config.roi_align_finest_scale,
mask=True)
self.roi_align_mask.set_train_local(config, True)
self.roi_align_test = SingleRoIExtractor(config,
config.roi_layer,
config.roi_align_out_channels,
config.roi_align_featmap_strides,
1,
config.roi_align_finest_scale,
mask=False)
self.roi_align_test.set_train_local(config, False)
self.roi_align_mask_test = SingleRoIExtractor(config,
config.roi_layer,
config.roi_align_out_channels,
config.roi_align_featmap_strides,
1,
config.roi_align_finest_scale,
mask=True)
self.roi_align_mask_test.set_train_local(config, False)
# Rcnn
self.rcnn_cls = RcnnCls(config, self.train_batch_size, self.num_classes)
self.rcnn_mask = RcnnMask(config, self.train_batch_size, self.num_classes)
# Op declare
self.squeeze = P.Squeeze()
self.cast = P.Cast()
self.concat = P.Concat(axis=0)
self.concat_1 = P.Concat(axis=1)
self.concat_2 = P.Concat(axis=2)
self.reshape = P.Reshape()
self.select = P.Select()
self.greater = P.Greater()
self.transpose = P.Transpose()
# Test mode
self.test_batch_size = config.test_batch_size
self.split = P.Split(axis=0, output_num=self.test_batch_size)
self.split_shape = P.Split(axis=0, output_num=4)
self.split_scores = P.Split(axis=1, output_num=self.num_classes)
self.split_fb_mask = P.Split(axis=1, output_num=self.num_classes)
self.split_cls = P.Split(axis=0, output_num=self.num_classes-1)
self.tile = P.Tile()
self.gather = P.GatherNd()
self.rpn_max_num = config.rpn_max_num
self.zeros_for_nms = Tensor(np.zeros((self.rpn_max_num, 3)).astype(np.float16))
self.ones_mask = np.ones((self.rpn_max_num, 1)).astype(np.bool)
self.zeros_mask = np.zeros((self.rpn_max_num, 1)).astype(np.bool)
self.bbox_mask = Tensor(np.concatenate((self.ones_mask, self.zeros_mask,
self.ones_mask, self.zeros_mask), axis=1))
self.nms_pad_mask = Tensor(np.concatenate((self.ones_mask, self.ones_mask,
self.ones_mask, self.ones_mask, self.zeros_mask), axis=1))
self.test_score_thresh = Tensor(np.ones((self.rpn_max_num, 1)).astype(np.float16) * config.test_score_thr)
self.test_score_zeros = Tensor(np.ones((self.rpn_max_num, 1)).astype(np.float16) * 0)
self.test_box_zeros = Tensor(np.ones((self.rpn_max_num, 4)).astype(np.float16) * -1)
self.test_iou_thr = Tensor(np.ones((self.rpn_max_num, 1)).astype(np.float16) * config.test_iou_thr)
self.test_max_per_img = config.test_max_per_img
self.nms_test = P.NMSWithMask(config.test_iou_thr)
self.softmax = P.Softmax(axis=1)
self.logicand = P.LogicalAnd()
self.oneslike = P.OnesLike()
self.test_topk = P.TopK(sorted=True)
self.test_num_proposal = self.test_batch_size * self.rpn_max_num
# Improve speed
self.concat_start = min(self.num_classes - 2, 55)
self.concat_end = (self.num_classes - 1)
# Init tensor
roi_align_index = [np.array(np.ones((config.num_expected_pos_stage2 + config.num_expected_neg_stage2, 1)) * i,
dtype=np.float16) for i in range(self.train_batch_size)]
roi_align_index_test = [np.array(np.ones((config.rpn_max_num, 1)) * i, dtype=np.float16) \
for i in range(self.test_batch_size)]
self.roi_align_index_tensor = Tensor(np.concatenate(roi_align_index))
self.roi_align_index_test_tensor = Tensor(np.concatenate(roi_align_index_test))
roi_align_index_pos = [np.array(np.ones((config.num_expected_pos_stage2, 1)) * i,
dtype=np.float16) for i in range(self.train_batch_size)]
self.roi_align_index_tensor_pos = Tensor(np.concatenate(roi_align_index_pos))
self.rcnn_loss_cls_weight = Tensor(np.array(config.rcnn_loss_cls_weight).astype(np.float16))
self.rcnn_loss_reg_weight = Tensor(np.array(config.rcnn_loss_reg_weight).astype(np.float16))
self.rcnn_loss_mask_fb_weight = Tensor(np.array(config.rcnn_loss_mask_fb_weight).astype(np.float16))
self.argmax_with_value = P.ArgMaxWithValue(axis=1)
self.on_value = Tensor(1.0, mstype.float32)
self.off_value = Tensor(0.0, mstype.float32)
self.onehot = P.OneHot()
self.reducesum = P.ReduceSum()
self.sigmoid = P.Sigmoid()
self.expand_dims = P.ExpandDims()
self.test_mask_fb_zeros = Tensor(np.zeros((self.rpn_max_num, 28, 28)).astype(np.float16))
self.value = Tensor(1.0, mstype.float16)
def _IgammacContinuedFraction(ax, x, a, enabled):
"""Helper function for computing Igammac using a continued fraction."""
abs_x = P.Abs()
logicaland = P.LogicalAnd()
greater = P.Greater()
less = P.Less()
notequal = P.NotEqual()
fill = P.Fill()
shape = P.Shape()
dtype = P.DType()
select = P.Select()
if dtype(ax) == mstype.float16:
epsilon = eps_fp16
else:
epsilon = eps_fp32
def cond(vals):
enabled = vals[0]
c = vals[5]
return logicaland(less(c, 2000), enabled)
def body(vals):
enabled = vals[0]
ans = vals[1]
t = vals[2]
y = vals[3]
z = vals[4]
c = vals[5]
pkm1 = vals[6]
qkm1 = vals[7]
pkm2 = vals[8]
qkm2 = vals[9]
dpkm2_da = vals[10]
dqkm2_da = vals[11]
dpkm1_da = vals[12]
dqkm1_da = vals[13]
dans_da = vals[14]
c = c + 1
y = y + 1
z = z + 2
yc = y * c
pk = pkm1 * z - pkm2 * yc
qk = qkm1 * z - qkm2 * yc
qk_is_nonzero = notequal(qk, 0)
r = pk / qk
t = select(qk_is_nonzero, abs_x((ans - r) / r), fill(dtype(t), shape(t), 1))
ans = select(qk_is_nonzero, r, ans)
dpk_da = dpkm1_da * z - pkm1 - dpkm2_da * yc + pkm2 * c
dqk_da = dqkm1_da * z - qkm1 - dqkm2_da * yc + qkm2 * c
dans_da_new = select(qk_is_nonzero, (dpk_da - ans * dqk_da) / qk, dans_da)
grad_conditional = select(qk_is_nonzero,
abs_x(dans_da_new - dans_da),
fill(dtype(dans_da), shape(dans_da), 1))
pkm2 = pkm1
pkm1 = pk
qkm2 = qkm1
qkm1 = qk
dpkm2_da = dpkm1_da
dqkm2_da = dqkm1_da
dpkm1_da = dpk_da
dqkm1_da = dqk_da
rescale = greater(abs_x(pk), 1 / epsilon)
pkm2 = select(rescale, pkm2 * epsilon, pkm2)
pkm1 = select(rescale, pkm1 * epsilon, pkm1)
qkm2 = select(rescale, qkm2 * epsilon, qkm2)
qkm1 = select(rescale, qkm1 * epsilon, qkm1)
dpkm2_da = select(rescale, dpkm2_da * epsilon, dpkm2_da)
dqkm2_da = select(rescale, dqkm2_da * epsilon, dqkm2_da)
dpkm1_da = select(rescale, dpkm1_da * epsilon, dpkm1_da)
dqkm1_da = select(rescale, dqkm1_da * epsilon, dqkm1_da)
conditional = logicaland(enabled, greater(grad_conditional, epsilon))
return (conditional, select(enabled, ans, vals[1]), select(enabled, t, vals[2]),
select(enabled, y, vals[3]), select(enabled, z, vals[4]),
c, select(enabled, pkm1, vals[6]),
select(enabled, qkm1, vals[7]), select(enabled, pkm2, vals[8]),
select(enabled, qkm2, vals[9]), select(enabled, dpkm2_da, vals[10]),
select(enabled, dqkm2_da, vals[11]), select(enabled, dpkm1_da, vals[12]),
select(enabled, dqkm1_da, vals[13]), select(enabled, dans_da_new, vals[14]))
y = 1 - a
z = x + y + 1
c = fill(dtype(x), shape(x), 0)
pkm2 = fill(dtype(x), shape(x), 1)
qkm2 = x
pkm1 = x + 1
qkm1 = z * x
ans = pkm1 / qkm1
t = fill(dtype(x), shape(x), 1)
dpkm2_da = fill(dtype(x), shape(x), 0)
dqkm2_da = fill(dtype(x), shape(x), 0)
dpkm1_da = fill(dtype(x), shape(x), 0)
dqkm1_da = -x
dans_da = (dpkm1_da - ans * dqkm1_da) / qkm1
vals = (enabled, ans, t, y, z, c, pkm1, qkm1, pkm2, qkm2, dpkm2_da, dqkm2_da, dpkm1_da, dqkm1_da, dans_da)
vals = _while_helper_func(cond, body, vals)
ans = vals[1]
return ans * ax
def __init__(self, config, batch_size, num_bboxes, add_gt_as_proposals):
super(BboxAssignSampleForRcnn, self).__init__()
cfg = config
self.dtype = np.float32
self.ms_type = mstype.float32
self.batch_size = batch_size
self.neg_iou_thr = cfg.neg_iou_thr_stage2
self.pos_iou_thr = cfg.pos_iou_thr_stage2
self.min_pos_iou = cfg.min_pos_iou_stage2
self.num_gts = cfg.num_gts
self.num_bboxes = num_bboxes
self.num_expected_pos = cfg.num_expected_pos_stage2
self.num_expected_neg = cfg.num_expected_neg_stage2
self.num_expected_total = cfg.num_expected_total_stage2
self.add_gt_as_proposals = add_gt_as_proposals
self.label_inds = Tensor(
np.arange(1, self.num_gts + 1).astype(np.int32))
self.add_gt_as_proposals_valid = Tensor(
np.full(self.num_gts, self.add_gt_as_proposals, dtype=np.int32))
self.concat = P.Concat(axis=0)
self.max_gt = P.ArgMaxWithValue(axis=0)
self.max_anchor = P.ArgMaxWithValue(axis=1)
self.sum_inds = P.ReduceSum()
self.iou = P.IOU()
self.greaterequal = P.GreaterEqual()
self.greater = P.Greater()
self.select = P.Select()
self.gatherND = P.GatherNd()
self.squeeze = P.Squeeze()
self.cast = P.Cast()
self.logicaland = P.LogicalAnd()
self.less = P.Less()
self.random_choice_with_mask_pos = P.RandomChoiceWithMask(
self.num_expected_pos)
self.random_choice_with_mask_neg = P.RandomChoiceWithMask(
self.num_expected_neg)
self.reshape = P.Reshape()
self.equal = P.Equal()
self.bounding_box_encode = P.BoundingBoxEncode(means=(0.0, 0.0, 0.0,
0.0),
stds=(0.1, 0.1, 0.2,
0.2))
self.concat_axis1 = P.Concat(axis=1)
self.logicalnot = P.LogicalNot()
self.tile = P.Tile()
# Check
self.check_gt_one = Tensor(
np.full((self.num_gts, 4), -1, dtype=self.dtype))
self.check_anchor_two = Tensor(
np.full((self.num_bboxes, 4), -2, dtype=self.dtype))
# Init tensor
self.assigned_gt_inds = Tensor(np.full(num_bboxes, -1, dtype=np.int32))
self.assigned_gt_zeros = Tensor(
np.array(np.zeros(num_bboxes), dtype=np.int32))
self.assigned_gt_ones = Tensor(
np.array(np.ones(num_bboxes), dtype=np.int32))
self.assigned_gt_ignores = Tensor(
np.full(num_bboxes, -1, dtype=np.int32))
self.assigned_pos_ones = Tensor(
np.array(np.ones(self.num_expected_pos), dtype=np.int32))
self.gt_ignores = Tensor(np.full(self.num_gts, -1, dtype=np.int32))
self.range_pos_size = Tensor(
np.arange(self.num_expected_pos).astype(self.dtype))
self.check_neg_mask = Tensor(
np.array(np.ones(self.num_expected_neg - self.num_expected_pos),
dtype=np.bool))
self.bboxs_neg_mask = Tensor(
np.zeros((self.num_expected_neg, 4), dtype=self.dtype))
self.labels_neg_mask = Tensor(
np.array(np.zeros(self.num_expected_neg), dtype=np.uint8))
self.reshape_shape_pos = (self.num_expected_pos, 1)
self.reshape_shape_neg = (self.num_expected_neg, 1)
self.scalar_zero = Tensor(0.0, dtype=self.ms_type)
self.scalar_neg_iou_thr = Tensor(self.neg_iou_thr, dtype=self.ms_type)
self.scalar_pos_iou_thr = Tensor(self.pos_iou_thr, dtype=self.ms_type)
self.scalar_min_pos_iou = Tensor(self.min_pos_iou, dtype=self.ms_type)
# shape of x and y not match
('LessEqual2', {
'block': (P.LessEqual(), {'exception': ValueError, 'error_keywords': ['LessEqual']}),
'desc_inputs': [Tensor(np.ones([3, 4]).astype(np.float32)), Tensor(np.ones([3, 2]).astype(np.float32))],
'skip': ['backward']}),
# input x is not Tensor(bool)
('LogicalNot1', {
'block': (P.LogicalNot(),
{'exception': TypeError, 'error_keywords': ['LogicalNot']}),
'desc_inputs': [Tensor(np.ones([2, 3]).astype(np.int32))],
'skip': ['backward']}),
# type of x and y not match
('LogicalAnd1', {
'block': (P.LogicalAnd(), {'exception': TypeError, 'error_keywords': ['LogicalAnd']}),
'desc_inputs': [Tensor(np.ones([3, 4]).astype(np.int32)), Tensor(np.ones([3, 4]).astype(np.bool_))],
'skip': ['backward']}),
# shape of x and y not match
('LogicalAnd2', {
'block': (P.LogicalAnd(), {'exception': ValueError, 'error_keywords': ['LogicalAnd']}),
'desc_inputs': [Tensor(np.ones([3, 4]).astype(np.bool_)), Tensor(np.ones([3, 2]).astype(np.bool_))],
'skip': ['backward']}),
# type of x and y not match
('LogicalOr1', {
'block': (P.LogicalOr(), {'exception': TypeError, 'error_keywords': ['LogicalOr']}),
'desc_inputs': [Tensor(np.ones([3, 4]).astype(np.int32)), Tensor(np.ones([3, 4]).astype(np.bool_))],
'skip': ['backward']}),
# shape of x and y not match
('LogicalOr2', {
def __init__(self, strategy1, strategy2):
super().__init__()
self.matmul = P.MatMul().shard(strategy1)
self.equal = P.Equal().shard(strategy2)
self.notequal = P.NotEqual().shard(strategy2)
self.logical = P.LogicalAnd().shard(strategy2)
