Align and crop a face image by landmarks. Arguments: img : Face image to be aligned and cropped. src_landmarks

Align and crop a face image by landmarks. Arguments: img : Face image to be aligned and cropped. src_landmarks

Parse args using the global parser.

Parse args using the global parser.

Load args from xxx ignoring type and choices with default still valid. Parameters ---------- parser: function Should return a dic

Load args from xxx ignoring type and choices with default still valid. Parameters ---------- parser: function Should return a dic

Wrap argparse.add_argument.

Wrap argparse.add_argument.

Calculate the average gradient for each shared variable across all towers. Note that this function provides a synchronization point across all to

Color jitter. Examples -------- >>> color_jitter(img, 25, 0.2, 0.2, 0.1)

Transform opencv images to [-1.0, 1.0].

Transform opencv images to [-1.0, 1.0].

Batch dataset of disk image for PNG and JPEG. Parameters ---------- img_paths : 1d-tensor/ndarray/list of str labels : nested structu

Return the current elapsed time since last (re)start.

Return the current elapsed time since last (re)start.

Transform images from [0.0, 1.0] to [-1.0, 1.0] of float64.

Transform images from [0.0, 1.0] to [-1.0, 1.0] of float64.

Gram-Schmidt process. Modified from https://stackoverflow.com/questions/48119473. Parameters ---------- vectors: 2D tensor - [v1, v2,

Merge the grid as an image.

Merge the grid as an image.

Split the image into a grid.

Split the image into a grid.

Transform images from [-1.0, 1.0] to opencv images.

Transform images from [-1.0, 1.0] to opencv images.

Transform images from [-1.0, 1.0] to [0.0, 1.0] of float64.

Transform images from [-1.0, 1.0] to [0.0, 1.0] of float64.

Crop an image with padding non-exisiting range. Parameters ---------- pad_mode: To be passed to skimage.util.pad as `mode` parame

Crop an image with padding non-exisiting range. Parameters ---------- pad_mode: To be passed to skimage.util.pad as `mode` parame

Merge images to an image with (n_rows * h) * (n_cols * w). Parameters ---------- images : numpy.array or object which can be converted to

Merge images to an image with (n_rows * h) * (n_cols * w). Parameters ---------- images : numpy.array or object which can be converted to

Return a float64 image in [-1.0, 1.0]. Parameters ---------- kwargs: To be passed to skimage.io.imread.

Show a [-1.0, 1.0] image.

Show a [-1.0, 1.0] image.

Save a [-1.0, 1.0] image. Parameters ---------- quality: Quality for JEPG format. plugin_args: To be passed to skimag

Run codes that are commonly used by LynnHo.