Init MegaFaceEvaluator by some initial files. Args: facescrub_json_list(str): the facescrub list provided by official.

Adjust model's first convolution layer to in_channels, if in_channels not equals 3. Args: in_channels (int): Input data's channel

Validates model name. Args: model_name (str): Name for efficientnet. Returns: bool: Is a valid name or not.

Forge transform maps with a given function `fn`. Args: output_shape (tuple): (b, h, w, ...). fn (Callable[[torch.Tensor], torch.T

Add mask to one image. Args: image_path(str): the image to add mask. face_lms(str): face landmarks, [x1, y1, x2, y2,

Get the index/indexes of the Nth largest element. Suppose the shape of score_mat is R*C, axis=1 mean that the indexes of the Nth larg

Get the index/indexes of the Nth largest element. Suppose the shape of score_mat is R*C, axis=1 mean that the indexes of the Nth larg

compute contrast depth in both of (out, label)

3x3 convolution with padding

3x3 convolution with padding

3x3 convolution with padding

3x3 convolution with padding

Should be overridden by all subclasses. Used for online image cropping, input the original Mat, and return the Mat obtained from the

Create BlockArgs and GlobalParams for efficientnet model. Args: width_coefficient (float) depth_coefficient (float) image

Map EfficientNet model name to parameter coefficients. Args: model_name (str): Model name to be queried. Returns: params_dic

Encode a list of BlockArgs to a list of strings. Args: blocks_args (list[namedtuples]): A list of BlockArgs namedtuples of block

Extract and return features. Args: model(object): initialized model. data_loader(object): load data to be

Extract and return features. Args: model(object): initialized model. data_loader(object): load data to be

Get the current learning rate from optimizer.

Judge whether the point is in the triangle Method: http://blackpawn.com/texts/pointinpoly/ Args: point: [u, v] or [x, y]

Update the gallery-set network in the momentum way.(MoCo)

Update the gallery-set network in the momentum way.(MoCo)

Remove the noise in facescrub. We use the class center of certain id as the feature of noise faces. Args: facescrub_noises_file(str):

Remove the noise in facescrub. We use the class center of certain id as the feature of noise faces. Args: facescrub_noises_file(str):

render mesh by z buffer Args: vertices: 3 x nver colors: 3 x nver triangles: 3 x ntri h: height w: width

ShuffleBN for batch, the same as MoCo https://arxiv.org/abs/1911.05722 #######

ShuffleBN for batch, the same as MoCo https://arxiv.org/abs/1911.05722 #######

Transform a image by cv2.

Args: x: (B, H, W, C) window_size (int): window size Returns: windows: (num_windows*B, window_size, window_size, C)

Init priorBox settings related to the generation of anchors.

Get a block through a string notation of arguments. Args: block_string (str): A string notation of arguments.

Encode a block to a string. Args: block (namedtuple): A BlockArgs type argument. Returns: block_string: A St