name: MaxRoiPool (GitHub)
shape inference: True
This version of the operator has been available since version 1.
ROI max pool consumes an input tensor X and region of interests (RoIs) to apply max pooling across each RoI, to produce output 4-D tensor of shape (num_rois, channels, pooled_shape, pooled_shape).
pooled_shape (required): ROI pool output shape (height, width).
spatial_scale: Multiplicative spatial scale factor to translate ROI coordinates from their input scale to the scale used when pooling. Default value is
X (heterogeneous) - T: Input data tensor from the previous operator; dimensions for image case are (N x C x H x W), where N is the batch size, C is the number of channels, and H and W are the height and the width of the data.
rois (heterogeneous) - T: RoIs (Regions of Interest) to pool over. Should be a 2-D tensor of shape (num_rois, 5) given as [[batch_id, x1, y1, x2, y2], …].
Y (heterogeneous) - T: RoI pooled output 4-D tensor of shape (num_rois, channels, pooled_shape, pooled_shape).
T in ( tensor(double), tensor(float), tensor(float16) ): Constrain input and output types to float tensors.