Python nd.pick方法代码示例

# 需要导入模块: from mxnet import nd [as 别名]
# 或者: from mxnet.nd import pick [as 别名]
def hybrid_forward(self, F, X, y=None):
        # import pdb; pdb.set_trace()
        X = self.net[0](X) # Conv1
        X = self.net[1](X) # Primary Capsule
        X = self.net[2](X) # Digital Capsule
        # import pdb ; pdb.set_trace()
        X = X.reshape((X.shape[0],X.shape[2], X.shape[4]))
        # get length of vector for margin loss calculation
        X_l2norm = nd.sqrt((X**2).sum(axis=-1))
        # import pdb ; pdb.set_trace()
        prob = nd.softmax(X_l2norm, axis=-1)

        if y is not None:
            max_len_indices = y
        else:
            
            max_len_indices = nd.argmax(prob,axis=-1)


        y_tile = nd.tile(y.expand_dims(axis=1), reps=(1, X.shape[-1]))
        batch_activated_capsules = nd.pick(X, y_tile, axis=1, keepdims=True)

        reconstrcutions = self.net[3](batch_activated_capsules)

        return  prob, X_l2norm, reconstrcutions 

# 需要导入模块: from mxnet import nd [as 别名]
# 或者: from mxnet.nd import pick [as 别名]
def hybrid_forward(self, F, pred, label):
        """Compute loss"""
        softmaxout = F.SoftmaxOutput(
            pred, label.astype(pred.dtype), ignore_label=self._ignore_label,
            multi_output=self._sparse_label,
            use_ignore=True, normalization='valid' if self._size_average else 'null')
        if self._sparse_label:
            loss = -F.pick(F.log(softmaxout), label, axis=1, keepdims=True)
        else:
            label = _reshape_like(F, label, pred)
            loss = -F.sum(F.log(softmaxout) * label, axis=-1, keepdims=True)
        loss = F.where(label.expand_dims(axis=1) == self._ignore_label,
                       F.zeros_like(loss), loss)
        return F.mean(loss, axis=self._batch_axis, exclude=True) 

# 需要导入模块: from mxnet import nd [as 别名]
# 或者: from mxnet.nd import pick [as 别名]
def _mixup_forward(self, F, pred, label1, label2, lam, sample_weight=None):
        if not self._from_logits:
            pred = F.log_softmax(pred, self._axis)
        if self._sparse_label:
            loss1 = -F.pick(pred, label1, axis=self._axis, keepdims=True)
            loss2 = -F.pick(pred, label2, axis=self._axis, keepdims=True)
            loss = lam * loss1 + (1 - lam) * loss2
        else:
            label1 = _reshape_like(F, label1, pred)
            label2 = _reshape_like(F, label2, pred)
            loss1 = -F.sum(pred*label1, axis=self._axis, keepdims=True)
            loss2 = -F.sum(pred*label2, axis=self._axis, keepdims=True)
            loss = lam * loss1 + (1 - lam) * loss2
        loss = _apply_weighting(F, loss, self._weight, sample_weight)
        return F.mean(loss, axis=self._batch_axis, exclude=True) 

# 需要导入模块: from mxnet import nd [as 别名]
# 或者: from mxnet.nd import pick [as 别名]
def hybrid_forward(self, F, pred, label):
        """Compute loss"""
        softmaxout = F.SoftmaxOutput(
            pred, label.astype(pred.dtype), ignore_label=self._ignore_label,
            multi_output=self._sparse_label,
            use_ignore=True, normalization='valid' if self._size_average else 'null')
        loss = -F.pick(F.log(softmaxout), label, axis=1, keepdims=True)
        loss = F.where(label.expand_dims(axis=1) == self._ignore_label,
                       F.zeros_like(loss), loss)
        return F.mean(loss, axis=self._batch_axis, exclude=True) 

# 需要导入模块: from mxnet import nd [as 别名]
# 或者: from mxnet.nd import pick [as 别名]
def _gather_feat(feat, ind, mask=None):
    # K cannot be 1 for this implementation
    K = ind.shape[1]
    batch_size = ind.shape[0]
    attri_dim = feat.shape[2]

    flatten_ind = ind.flatten()
    for i in range(batch_size):
        if i == 0:
            output = feat[i, ind[i]].expand_dims(2)   # similar to nd.pick
        else:
            output = nd.concat(output, feat[i, ind[i]].expand_dims(2), dim=2)
    output = output.swapaxes(dim1 = 1, dim2 = 2)
    return output 

# 需要导入模块: from mxnet import nd [as 别名]
# 或者: from mxnet.nd import pick [as 别名]
def forward(self, cls_pred, box_pred, cls_target, box_target):
        """Compute loss in entire batch across devices."""
        # require results across different devices at this time
        cls_pred, box_pred, cls_target, box_target = [_as_list(x) \
            for x in (cls_pred, box_pred, cls_target, box_target)]
        # cross device reduction to obtain positive samples in entire batch
        num_pos = []
        for cp, bp, ct, bt in zip(*[cls_pred, box_pred, cls_target, box_target]):
            pos_samples = (ct > 0)
            num_pos.append(pos_samples.sum())
        num_pos_all = sum([p.asscalar() for p in num_pos])
        if num_pos_all < 1:
            # no positive samples found, return dummy losses
            return nd.zeros((1,)), nd.zeros((1,)), nd.zeros((1,))

        # compute element-wise cross entropy loss and sort, then perform negative mining
        cls_losses = []
        box_losses = []
        sum_losses = []
        for cp, bp, ct, bt in zip(*[cls_pred, box_pred, cls_target, box_target]):
            pred = nd.log_softmax(cp, axis=-1)
            pos = ct > 0
            cls_loss = -nd.pick(pred, ct, axis=-1, keepdims=False)
            rank = (cls_loss * (pos - 1)).argsort(axis=1).argsort(axis=1)
            hard_negative = rank < (pos.sum(axis=1) * self._negative_mining_ratio).expand_dims(-1)
            # mask out if not positive or negative
            cls_loss = nd.where((pos + hard_negative) > 0, cls_loss, nd.zeros_like(cls_loss))
            cls_losses.append(nd.sum(cls_loss, axis=0, exclude=True) / num_pos_all)

            bp = _reshape_like(nd, bp, bt)
            box_loss = nd.abs(bp - bt)
            box_loss = nd.where(box_loss > self._rho, box_loss - 0.5 * self._rho,
                                (0.5 / self._rho) * nd.square(box_loss))
            # box loss only apply to positive samples
            box_loss = box_loss * pos.expand_dims(axis=-1)
            box_losses.append(nd.sum(box_loss, axis=0, exclude=True) / num_pos_all)
            sum_losses.append(cls_losses[-1] + self._lambd * box_losses[-1])

        return sum_losses, cls_losses, box_losses 

# 需要导入模块: from mxnet import nd [as 别名]
# 或者: from mxnet.nd import pick [as 别名]
def forward(self, cls_pred, box_pred, cls_target, box_target):
        """Compute loss in entire batch across devices."""
        # require results across different devices at this time
        cls_pred, box_pred, cls_target, box_target = [_as_list(x) \
            for x in (cls_pred, box_pred, cls_target, box_target)]
        # cross device reduction to obtain positive samples in entire batch
        num_pos = []
        for cp, bp, ct, bt in zip(*[cls_pred, box_pred, cls_target, box_target]):
            pos_samples = (ct > 0)
            num_pos.append(pos_samples.sum())
        num_pos_all = sum([p.asscalar() for p in num_pos])
        if num_pos_all < 1 and self._min_hard_negatives < 1:
            # no positive samples and no hard negatives, return dummy losses
            cls_losses = [nd.sum(cp * 0) for cp in cls_pred]
            box_losses = [nd.sum(bp * 0) for bp in box_pred]
            sum_losses = [nd.sum(cp * 0) + nd.sum(bp * 0) for cp, bp in zip(cls_pred, box_pred)]
            return sum_losses, cls_losses, box_losses


        # compute element-wise cross entropy loss and sort, then perform negative mining
        cls_losses = []
        box_losses = []
        sum_losses = []
        for cp, bp, ct, bt in zip(*[cls_pred, box_pred, cls_target, box_target]):
            pred = nd.log_softmax(cp, axis=-1)
            pos = ct > 0
            cls_loss = -nd.pick(pred, ct, axis=-1, keepdims=False)
            rank = (cls_loss * (pos - 1)).argsort(axis=1).argsort(axis=1)
            hard_negative = rank < nd.maximum(self._min_hard_negatives, pos.sum(axis=1)
                                              * self._negative_mining_ratio).expand_dims(-1)
            # mask out if not positive or negative
            cls_loss = nd.where((pos + hard_negative) > 0, cls_loss, nd.zeros_like(cls_loss))
            cls_losses.append(nd.sum(cls_loss, axis=0, exclude=True) / max(1., num_pos_all))

            bp = _reshape_like(nd, bp, bt)
            box_loss = nd.abs(bp - bt)
            box_loss = nd.where(box_loss > self._rho, box_loss - 0.5 * self._rho,
                                (0.5 / self._rho) * nd.square(box_loss))
            # box loss only apply to positive samples
            box_loss = box_loss * pos.expand_dims(axis=-1)
            box_losses.append(nd.sum(box_loss, axis=0, exclude=True) / max(1., num_pos_all))
            sum_losses.append(cls_losses[-1] + self._lambd * box_losses[-1])

        return sum_losses, cls_losses, box_losses