Skip to content

Instantly share code, notes, and snippets.

@dudeperf3ct
dudeperf3ct / export_tflite.py
Created August 2, 2022 14:09
Export to TFLite
from functools import partial
import numpy as np
import os
import glob
import argparse
import tensorflow as tf
from tensorflow.keras.preprocessing.image import ImageDataGenerator
from config.config import logger
"""
Script to augment teaching data
Usage:
augment.py --path=<source-dir> [--out=<target_dir>] --aug=<name> [--prob=<PROBABILITY>] [--add-prefix] [--start-id=<id>]
Options:
-h --help Show this screen.
--path TUBPATH Path of the record directory
--out TARGETPATH Path to directory where augmented data is placed, directory is created if it doesn't exist
If omitted path (source directory) will be used
--aug NAME Augmentation type. Possible values: shadow, bright, flip