CheckBinaryVersion */

CheckBinaryVersion */

Extract schema from a given class Args: cls (type): Class from which to extract. Returns: schema (SchemaDict): Extracte

gt dir: (wider_face_val.mat, wider_easy_val.mat, wider_medium_val.mat, wider_hard_val.mat)

gt dir: (wider_face_val.mat, wider_easy_val.mat, wider_medium_val.mat, wider_hard_val.mat)

single image evaluation pred: Nx5 gt: Nx4 ignore:

single image evaluation pred: Nx5 gt: Nx4 ignore:

We resize both tensors to [A,B,2] without new malloc: [A,2] -> [A,1,2] -> [A,B,2] [B,2] -> [1,B,2] -> [A,B,2] Then we compute the area of

return [[img_name, num_face, [x0, y0, w, h], [...], [...]...], ...] TODO: the meaning of other bool value in input_txt, e.g. is_blur

Merge config into global config. Args: config (dict): Config to be merged. Returns: global config

norm score pred {key: [[x1,y1,x2,y2,s]]}

norm score pred {key: [[x1,y1,x2,y2,s]]}

ret: [[img_name(not abs_path), dir_name(not abs_path)], ...]

Constructs a ResNet-18 model. Args: pretrained (bool): If True, returns a model pre-trained on ImageNet progress (bool): If True,

Compute the IoU of two sets of boxes. E.g.: A ∩ B / A = A ∩ B / area(A) Args: box_a: (tensor) Ground truth bounding boxes,

Match each prior box with the ground truth box of the highest jaccard overlap, encode the bounding boxes, then return the matched indices corr

Match each prior box with the ground truth box of the highest jaccard overlap, encode the bounding boxes, then return the matched indices corr