import matplotlib import cv2 from tensorflow.keras.preprocessing.image import ImageDataGenerator import matplotlib.pyplot as plt import numpy as np import argparse import pickle import os import math import tensorflow as tf import Fetcher import Model import Config from tqdm.autonotebook import tqdm fetcher = Fetcher.DataFetcher(); #fetcher.processAndSave(); #fetcher.loadGestures(gestures= ["1","2"]); fetcher.load(); model = Model.Net(numBatchTrain = 0); model.network.summary(); model.load_checkpoint(); numSamples = 10; for i in range(numSamples): image, labels = fetcher.getRandomValidation();
from Models import * import tensorflow as tf import Config import pickle import Fetcher import cv2 import math from matplotlib import pyplot as pp import pydot from tqdm.autonotebook import tqdm tf.config.optimizer.set_jit(True) data = Fetcher.DataFetcher(imagePath="Dataset/fundus", heightmapPath="Dataset/heightmap") data.load() batch_count_train = int(np.ceil(data.trainDataSize / Config.BATCH_SIZE)) batch_count_valid = int(np.ceil(data.validDataSize / Config.BATCH_SIZE)) batch_count_test = int(np.ceil(data.testDataSize / Config.BATCH_SIZE)) gen_to_disc_ratio = 3 net = FundusNet(numBatchTrain=int(batch_count_train) * gen_to_disc_ratio) net.generator.summary() net.generator.load_weights(filepath="ckpt_weights1\\weights", by_name=True) startEpoch = net.load_checkpoint() # Define loss and accuracy per epoch all_loss = np.zeros(batch_count_train, dtype='float32')