training works

config2.py

@@ -0,0 +1,9 @@
+import os
+from utils import *
+
+data_dir = os.path.join(os.path.dirname(__file__), 'data')
+
+virat = new(dir = os.path.join(data_dir, 'VIRAT/Public Dataset/VIRAT Video Dataset Release 2.0/'))
+virat.aerial = new(dir = os.path.join(virat.dir, 'VIRAT Aerial Dataset'))
+virat.ground = new(dir = os.path.join(virat.dir, 'VIRAT Ground Dataset'))
+virat.ground.video = new(dir = os.path.join(virat.ground.dir, 'videos_original'))

srgan/train.py

@@ -1,5 +1,6 @@
 #! /usr/bin/python
 # -*- coding: utf8 -*-
+#%%
 
 import os
 import time
@@ -11,6 +12,12 @@
 from model import get_G, get_D
 from config import config
 
+import sys
+sys.path.append('..')
+import videodataset
+from config2 import *
+import glob2
+
 ###====================== HYPER-PARAMETERS ===========================###
 ## Adam
 batch_size = config.TRAIN.batch_size  # use 8 if your GPU memory is small, and change [4, 4] in tl.vis.save_images to [2, 4]
@@ -34,13 +41,13 @@
 
 def get_train_data():
 	# load dataset
-	train_hr_img_list = sorted(tl.files.load_file_list(path=config.TRAIN.hr_img_path, regx='.*.png', printable=False))#[0:20]
+	# train_hr_img_list = sorted(tl.files.load_file_list(path=config.TRAIN.hr_img_path, regx='.*.png', printable=False))#[0:20]
 		# train_lr_img_list = sorted(tl.files.load_file_list(path=config.TRAIN.lr_img_path, regx='.*.png', printable=False))
 		# valid_hr_img_list = sorted(tl.files.load_file_list(path=config.VALID.hr_img_path, regx='.*.png', printable=False))
 		# valid_lr_img_list = sorted(tl.files.load_file_list(path=config.VALID.lr_img_path, regx='.*.png', printable=False))
 
 	## If your machine have enough memory, please pre-load the entire train set.
-	train_hr_imgs = tl.vis.read_images(train_hr_img_list, path=config.TRAIN.hr_img_path, n_threads=32)
+	# train_hr_imgs = tl.vis.read_images(train_hr_img_list, path=config.TRAIN.hr_img_path, n_threads=32)
 		# for im in train_hr_imgs:
 		#     print(im.shape)
 		# valid_lr_imgs = tl.vis.read_images(valid_lr_img_list, path=config.VALID.lr_img_path, n_threads=32)
@@ -51,9 +58,15 @@ def get_train_data():
 		#     print(im.shape)
 
 	# dataset API and augmentation
-	def generator_train():
-		for img in train_hr_imgs:
-			yield img
+	# def generator_train():
+	# 	for img in train_hr_imgs:
+	# 		yield img
+
+	videoPaths = np.array(glob2.glob(virat.ground.video.dir + '/*.mp4'))
+	generator = videodataset.FrameGenerator(videoPaths)
+
+
+
 	def _map_fn_train(img):
 		hr_patch = tf.image.random_crop(img, [384, 384, 3])
 		hr_patch = hr_patch / (255. / 2.)
@@ -63,7 +76,10 @@ def _map_fn_train(img):
 		return lr_patch, hr_patch
 
 
-	train_ds = tf.data.Dataset.from_generator(generator_train, output_types=(tf.float32))
+	train_ds = tf.data.Dataset.from_generator(generator.call, output_types=(tf.float32))
+	# train_ds = tf.data.Dataset.from_generator(generator_train, output_types=(tf.float32))
+	# print(next(iter(train_ds)).numpy())
+	# return
 	train_ds = train_ds.map(_map_fn_train, num_parallel_calls=multiprocessing.cpu_count())
 		# train_ds = train_ds.repeat(n_epoch_init + n_epoch)
 	train_ds = train_ds.shuffle(shuffle_buffer_size)

videodataset.py

@@ -34,11 +34,12 @@ def call(self):
 			idx_frame = np.random.choice(self.totalFrames[i_vid], self.totalFrames[i_vid], replace=True)
 			for i_frame in idx_frame:
 				vid.set(cv2.CAP_PROP_POS_FRAMES, i_frame) # set video to this frame
-				yield {
-					'video_index': i_vid,
-					'video_path': self.videoPaths[i_vid],
-					'frame': vid.read()[1]
-				}
+				# yield {
+				# 	'video_index': i_vid,
+				# 	'video_path': self.videoPaths[i_vid],
+				# 	'frame': vid.read()[1]
+				# }
+				yield vid.read()[1]
 
 if __name__ == "__main__":
 	videoPaths = np.array(glob2.glob(virat.ground.video.dir + '/*.mp4'))