Get the current learning rate from optimizer.

Get the current learning rate from optimizer.

Get the current learning rate from optimizer.

Get the current learning rate from optimizer.

Get the current learning rate from optimizer.

Get the feature list of masked facescrub. Returns: A numpy array, the shape is N * feat_dim, N is the number of images in facescr

Get the feature list of masked facescrub. Returns: A numpy array, the shape is N * feat_dim, N is the number of images in facescr

Get the feature list of each id in masked facescrub. Returns: A dict，key is the id in masked facescrub，value is the feature list

Get the feature list of each id in masked facescrub. Returns: A dict，key is the id in masked facescrub，value is the feature list

Get the feature list of megaface. Returns: A numpy array, the shape is N * feat_dim, N is the number of images in megaface.

Get the feature list of megaface. Returns: A numpy array, the shape is N * feat_dim, N is the number of images in megaface.

Get the block args and global params for a given model name. Args: model_name (str): Model's name. override_params (dict): A dict

Get new texture Mainly for data augmentation.

Get the weights of the position Methods: https://gamedev.stackexchange.com/questions/23743/whats-the-most-efficient-way-to-find-barycentric-coord

Get the rank N accuracy Args: cur_rank(int): current rank, cur_rank=1 for rank1 accuracy. Returns: accuracy(

Get the rank N accuracy Args: cur_rank(int): current rank, cur_rank=1 for rank1 accuracy. Returns: accuracy(

Args: from_pts, to_pts: b x n x 2 Returns: torch.Tensor: b x 3 x 3

Get vertices Args: face_lms: face landmarks. image:[0, 255]

Obtain height and width from x. Args: x (int, tuple or list): Data size. Returns: size: A tuple or list (H,W).

Load features to memory. Returns: image_name2feature(dict): key is the name of image, value is feature of image.

Load features to memory. Returns: image_name2feature(dict): key is the name of image, value is feature of image.

Loads pretrained weights from weights path or download using url. Args: model (Module): The whole model of efficientnet. model_na

The method for parsing pair list.

The method for parsing pair list.

Python version NMS. Returns: The kept index after NMS.

Keep reading image until succeed. This can avoid IOError incurred by heavy IO process.

Keep reading image until succeed. This can avoid IOError incurred by heavy IO process.

Remove the noise in megaface. We set the feature of noise faces to zero vector, since we use cos similarity as the distance metric. Args

Remove the noise in megaface. We set the feature of noise faces to zero vector, since we use cos similarity as the distance metric. Args

Constructs a ResNet-50 model.

Constructs a ResNet-50 model. Args: pretrained (bool): If True, returns a model pre-trained on ImageNet

Calculate module's repeat number of a block based on depth multiplier. Use depth_coefficient of global_params. Args: repeats (int)

Selects a single box from multiple for a given image using one of multiple heuristics. Arguments: all_boxes {np.ndarray} -- I

Sets swish function as memory efficient (for training) or standard (for export). Args: memory_efficient (bool): Whether to use me

Get the cmc curve for megaface Compute the accuracy of rank 1 to rank max_rank Args: max_rank(int): the max rank to comp

Get the cmc curve for megaface Compute the accuracy of rank 1 to rank max_rank Args: max_rank(int): the max rank to comp

Get the accuracy of a model. Args: test_pair_list(list): the pair list given by PairsParser. image_name2feat

Get the accuracy of a model. Args: test_pair_list(list): the pair list given by PairsParser. image_name2feat

Total training procedure.