def main():
"""main"""
# parse arguments
args = parse_args()
if args is None:
exit()
# open session
config = tf.ConfigProto(allow_soft_placement=True)
config.gpu_options.allow_growth = True
with tf.Session(config=config) as sess:
gan = UGATIT(sess, args)
# build graph
gan.build_model()
# show network architecture
show_all_variables()
if args.phase == 'train':
gan.train()
print(" [*] Training finished!")
if args.phase == 'test':
gan.test()
print(" [*] Test finished!")
if args.phase == "export":
gan.export_saved_model()
print(" [*] Export model finished!")
def selfie2anime():
img_id = os.environ['id']
result_id = os.environ['result']
parser = get_parser()
args = parser.parse_args("--phase test".split())
with tf.Session(config=tf.ConfigProto(allow_soft_placement=True)) as sess:
#sess.reuse_variables()
gan = UGATIT(sess, args)
# build graph
gan.build_model()
# download target img
download_path = os.path.join(img_path, img_id)
download_image(images_bucket, img_id, dest=download_path)
dataset_tool.create_from_images(record_path, img_path, True)
# os.remove(del_record)
img = gan.infer(download_path)
image_url = upload_image(img, result_id)
return download_path, img
def main():
# parse arguments
args = parse_args()
if args is None:
exit()
# open session
with tf.Session(config=tf.ConfigProto(allow_soft_placement=True)) as sess:
gan = UGATIT(sess, args)
# build graph
gan.build_model()
# show network architecture
show_all_variables()
if args.phase == 'train':
gan.train()
print(" [*] Training finished!")
if args.phase == 'test':
gan.test()
print(" [*] Test finished!")
if args.phase == 'export':
gan.export()
print(" [*] Export finished!")
def main():
# parse arguments
args = parse_args()
if args is None:
exit()
use_gpu = args.device == 'cuda'
place = fluid.CUDAPlace(0) if use_gpu else fluid.CPUPlace()
args.place = place
with fluid.dygraph.guard(place):
# open session
gan = UGATIT(args)
# build graph
gan.build_model()
if args.phase == 'train':
gan.train()
print(" [*] Training finished!")
if args.phase == 'test':
gan.test()
print(" [*] Test finished!")
if args.phase == 'deploy':
gan.deploy()
print(" [*] Deploy finished!")
def main():
# parse arguments
args = parse_args()
if args is None:
exit()
# 自定义参数
args.epoch = 1000
args.batch_size = 1
# args.dataset = "s2a4zhengsheng"
# args.gan_type = "dragan"
# open session
with tf.Session(config=tf.ConfigProto(allow_soft_placement=True)) as sess:
gan = UGATIT(sess, args)
# build graph
gan.build_model()
# show network architecture
show_all_variables()
if args.phase == 'train':
gan.train()
print(" [*] Training finished!")
if args.phase == 'test':
gan.test()
print(" [*] Test finished!")
def main():
# parse arguments
args = parse_args()
if args is None:
exit()
gpu_options = tf.GPUOptions(allow_growth=True)
# open session
with tf.Session(config=tf.ConfigProto(allow_soft_placement=True, gpu_options=gpu_options)) as sess:
gan = UGATIT(sess, args)
print('[Info] 构建模型开始!')
# build graph
gan.build_model()
print('[Info] 构建模型完成!')
# show network architecture
show_all_variables()
if args.phase == 'train':
gan.train()
print(" [*] Training finished!")
if args.phase == 'test':
gan.test()
print(" [*] Test finished!")
def main():
# parse arguments
args = parse_args()
if args is None:
exit()
# open session
with tf.Session(config=tf.ConfigProto(allow_soft_placement=True)) as sess:
gan = UGATIT(sess, args)
# build graph
gan.build_model()
# show network architecture
show_all_variables()
if args.phase == 'train':
gan.train()
print(" [*] Training finished!")
if args.phase == 'test':
# gan.test()
gan.test_single_img(
sample_file="UGATIT/dataset/selfie2anime/testA/conv.JPG"
) #pass the img you want to convert
print(" [*] Test finished!")
def main():
# parse arguments
# import os
# os.environ["CUDA_VISIBLE_DEVICES"] = "2"
args = parse_args()
import os
gpu_id = args.gpu_ids
os.environ["CUDA_VISIBLE_DEVICES"] = str(gpu_id)
if args is None:
exit()
# open session
gan = UGATIT(args)
# build graph
gan.build_model()
if args.phase == 'train':
gan.train()
print(" [*] Training finished!")
if args.phase == 'test':
gan.test()
print(" [*] Test finished!")
def main():
# parse arguments
args = parse_args()
if args is None:
exit()
from paddle import fluid
if args.device == "cuda":
place = fluid.CUDAPlace(fluid.dygraph.parallel.Env().dev_id)
else:
place = fluid.CPUPlace()
#
with fluid.dygraph.guard(place=place):
if args.parallel:
args.strategy = fluid.dygraph.parallel.prepare_context()
else:
args.strategy = None
gan = UGATIT(args)
# build graph
gan.build_model()
if args.phase == 'train':
gan.train()
print(" [*] Training finished!")
if args.phase == 'test':
gan.test()
print(" [*] Test finished!")
def main():
# parse arguments
args = parse_args()
if args is None:
exit()
if args.weather == "Day2rain":
args.dataset = "day2rain"
elif args.weather == "Day2night":
args.dataset = "day2night"
# open session
gan = UGATIT(args)
# build graph
gan.build_model()
if args.phase == 'train':
gan.train()
print(" [*] Training finished!")
if args.phase == 'test':
gan.generate(args.dataset_path, args.output_path)
print(" [*] Test finished!")
def process():
input_path = generate_random_filename(upload_directory, "jpg")
output_path = generate_random_filename(result_directory, "jpg")
try:
url = request.json["url"]
download(url, input_path)
with tf.Session(config=tf.ConfigProto(
allow_soft_placement=True)) as sess:
gan = UGATIT(sess, args)
gan.build_model()
gan.test_endpoint_init()
gan.test_endpoint(input_path, output_path)
callback = send_file(output_path, mimetype='image/jpeg')
return callback, 200
except:
traceback.print_exc()
return {'message': 'input error'}, 400
finally:
clean_all([input_path, output_path])
def index():
# set session for image results
if "file_urls" not in session:
session['file_urls'] = []
# list to hold our uploaded image urls
file_urls = session['file_urls']
if request.method == 'POST':
file_obj = request.files
for f in file_obj:
file = request.files.get(f)
# convert string of image data to uint8
nparr = np.fromfile(file, np.uint8)
# decode image
img = cv2.imdecode(nparr, cv2.IMREAD_COLOR)
# parse arguments
args = parse_args()
if args is None:
exit()
# open session
with tf.Session(config=tf.ConfigProto(
allow_soft_placement=True)) as sess:
gan = UGATIT(sess, args)
# build graph
gan.build_model()
# show network architecture
show_all_variables()
# do some fancy processing here....
fake_img = gan.test_endpoint(img)
# save the file with to our photos folder
filename = str(uuid.uuid1()) + '.png'
cv2.imwrite('uploads/' + filename, fake_img)
# append image urls
file_urls.append(photos.url(filename))
session['file_urls'] = file_urls
return "uploading..."
return render_template('index.html')
def video(args):
# open session
with tf.Session(config=tf.ConfigProto(allow_soft_placement=True)) as sess:
gan = UGATIT(sess, args)
# build graph
gan.build_model()
# show network architecture
show_all_variables()
# load model for use in video stream
gan.video_inference_init()
# Run frame thread
threading.Thread(target=read_frame_thread).start()
while state.running:
if state.frame is None:
time.sleep(0.01)
continue
# Get recent frame
frame = state.frame
# generate image
gen_image = gan.video_inference(frame)
# display frame
cv2.imshow('Output', gen_image)
# handle key press events
process_events()
def getUGATITTransform(input_image, dataset):
image = imageio.imread(io.BytesIO(input_image))
[h,w,c] = np.shape(image)
scale = 1
if((h > 512) or (w > 512)):
if (h > w):
scale = h // 512 + 1
else:
scale = w // 512 + 1
image = cv2.resize(image, dsize=(int(w/scale), int(h/scale)))
t_image = torch.from_numpy(np.expand_dims(np.transpose(image, (2, 0, 1)), 0)).cuda().float()
args = parse_args()
args.light = True
args.dataset=dataset
gan = UGATIT(args)
gan.build_model()
model_list = glob(os.path.join(gan.result_dir, gan.dataset, 'model', '*.pt'))
if not len(model_list) == 0:
model_list.sort()
iter = int(model_list[-1].split('_')[-1].split('.')[0])
gan.load(os.path.join(gan.result_dir, gan.dataset, 'model'), iter)
print(" [*] Load SUCCESS")
else:
print(" [*] Load FAILURE")
return
gan.genA2B.eval(), gan.genB2A.eval()
real_A = t_image
fake_A2B, _, fake_A2B_heatmap = gan.genA2B(real_A)
result = np.transpose(np.squeeze(fake_A2B.data.cpu().numpy()), (1,2,0))
return result
def main(args):
h, w = _FRAME_SIZE
x = tf.placeholder(tf.float32, [None, h, w, 3])
model = UGATIT()
y = model.generate(x)
with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
restorer = tf.train.Saver()
restorer.restore(sess, args.checkpoint)
cap = cv2.VideoCapture(_DEFAULT_CAMERA_ID)
if not cap.isOpened():
raise Exception("Unable to read camera feed")
# by default use timestamp to name recorded videos
video_path = os.path.join(
_DEFAULT_OUTPUT_DIR,
'{}.mp4'.format(datetime.now().strftime("%Y-%m-%d-%H%M%S")))
fourcc = cv2.VideoWriter_fourcc(*_DEFAULT_CODEC)
video_out = cv2.VideoWriter(
filename=video_path,
fourcc=fourcc,
fps=_DEFAULT_FRAMES_PER_SECOND,
frameSize=(w * 2, h) if args.sidebyside else _FRAME_SIZE)
window_message = "Press [SPACE BAR] to stop"
while cap.isOpened():
ret, orig = cap.read()
if ret:
orig = cv2.resize(orig, _FRAME_SIZE).astype(np.float32)
out = cv2.cvtColor(orig, cv2.COLOR_BGR2RGB) / 127.5 - 1.
out = sess.run(y, feed_dict={x: np.expand_dims(out, 0)})
out = (np.squeeze(out) + 1.) / 2.
out = cv2.cvtColor(out, cv2.COLOR_RGB2BGR)
frame = np.hstack(
(orig / 255., out)) if args.sidebyside else out
cv2.imshow(window_message, frame)
if args.save:
video_out.write((frame * 255).astype(np.uint8))
if cv2.waitKey(1) & 0xFF == ord(" "):
break
else:
break
video_out.release()
cap.release()
cv2.destroyAllWindows()
def main():
# parse arguments
args = parse_args()
# Change to fullscreen
cv2.namedWindow('Output', cv2.WND_PROP_FULLSCREEN)
cv2.setWindowProperty('Output', cv2.WND_PROP_FULLSCREEN,
cv2.WINDOW_FULLSCREEN)
if args is None:
exit()
if args.phase == 'video':
video(args)
# open session
with tf.Session(config=tf.ConfigProto(allow_soft_placement=True)) as sess:
gan = UGATIT(sess, args)
# build graph
gan.build_model()
# show network architecture
show_all_variables()
if args.phase == 'web':
gan.video_inference_init()
global gan_ref
gan_ref = gan
app.run(host="0.0.0.0", port=5000)
exit()
if args.phase == 'train':
gan.train()
print(" [*] Training finished!")
if args.phase == 'test':
gan.test()
print(" [*] Test finished!")
def main():
config = args()
# open session
gan = UGATIT(config)
# build graph
gan.build_model()
gan.train()
def setup(opts):
args = parse_args()
args.dataset = 'portrait'
gan = UGATIT(sess, args)
gan.build_model()
gan.load_from_latest(opts['checkpoint'])
return gan
def setup(opts):
args = parse_args()
args.phase = 'test'
args.img_size = 256
gan = UGATIT(sess, args)
gan.build_model()
gan.load_from_latest(opts['checkpoint'])
return gan
def main(args):
if not os.path.exists(args.input):
raise FileNotFoundError("Input image does not exist")
h, w = _IMAGE_SIZE
x = tf.placeholder(tf.float32, [None, h, w, 3])
model = UGATIT()
y = model.generate(x)
with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
restorer = tf.train.Saver()
restorer.restore(sess, args.checkpoint)
orig = cv2.imread(args.input)
orig = cv2.resize(orig, _IMAGE_SIZE).astype(np.float32)
out = cv2.cvtColor(orig, cv2.COLOR_BGR2RGB) / 127.5 - 1.
out = sess.run(y, feed_dict={x: np.expand_dims(out, 0)})
out = (np.squeeze(out) + 1.) / 2.
out = cv2.cvtColor(out, cv2.COLOR_RGB2BGR)
window_message = "Press any key to close"
if args.sidebyside:
cv2.imshow(window_message, np.hstack((orig / 255., out)))
else:
cv2.imshow(window_message, out)
cv2.waitKey(0)
cv2.destroyAllWindows()
if args.save:
base_filename = os.path.basename(args.input).split(".")[0]
out_path = os.path.join(_DEFAULT_OUTPUT_DIR,
base_filename + "_out.png")
save_out = (out * 255).astype(np.uint8)
cv2.imwrite(out_path, save_out)
print("output image saved to: {}".format(out_path))
if args.sidebyside: # also save side-by-side image
sbs_path = os.path.join(_DEFAULT_OUTPUT_DIR,
base_filename + "_sbs.png")
cv2.imwrite(sbs_path,
np.hstack((orig.astype(np.uint8), save_out)))
print("side-by-side image saved to: {}".format(sbs_path))
def main():
# parse arguments
args = parse_args()
if args is None:
exit()
# open session
gan = UGATIT(args)
# build graph
gan.build_model()
if args.phase == 'train':
gan.train()
print(" [*] Training finished!")
if args.phase == 'test':
gan.test()
print(" [*] Test finished!")
if args.phase == 'val':
gan.val()
print(" [*] Val finished!")
def main():
# parse arguments
args = parse_args()
if args is None:
exit()
if args.phase == 'video':
video(args)
# open session
with tf.Session(config=tf.ConfigProto(allow_soft_placement=True)) as sess:
gan = UGATIT(sess, args)
# build graph
gan.build_model()
# show network architecture
show_all_variables()
if args.phase == 'web':
gan.video_inference_init()
global gan_ref
gan_ref = gan
app.run(host="0.0.0.0", port=5000)
exit()
if args.phase == 'train':
gan.train()
print(" [*] Training finished!")
if args.phase == 'test':
gan.test()
print(" [*] Test finished!")
def main():
# parse arguments
args = parse_args()
if args is None:
exit()
gan = UGATIT(args)
# build graph
gan.build()
if args.phase == 'train':
gan.train()
print(" [*] Training finished!")
if args.phase == 'test':
gan.test()
print(" [*] Test finished!")
def main():
in_dir = os.getenv("INPUT_DIR")
out_dir = os.getenv("RESULT_DIR")
with tf.Session(config=tf.ConfigProto(allow_soft_placement=True)) as sess:
args = FakeArgs()
gan = UGATIT(sess, args)
gan.build_model()
tf.global_variables_initializer().run()
gan.loop_on_input(in_dir, out_dir)
def runner(args):
# open session
with tf.Session(config=tf.ConfigProto(allow_soft_placement=True)) as sess:
gan = UGATIT(sess, args)
# build graph
gan.build_model()
# show network architecture
show_all_variables()
gan.test_endpoint_init()
def load_resources(mode):
global face_classifier_classifier
face_classifier_classifier = cv2.CascadeClassifier(face_cascade_path)
if mode == modes.ANIME_MODE:
global anime_session, anime_model
anime_session = tf.Session(config=tf.ConfigProto(
allow_soft_placement=True))
anime_model = UGATIT(anime_session, args)
anime_model.build_model()
anime_model.load_model(anime_session)
def main():
# parse arguments
args = parse_args()
if args is None:
exit()
# open session
gan = UGATIT(args)
# build graph
gan.build_model()
if args.phase == 'train':
gan.train()
print(" [*] Training finished!")
if args.phase == 'test':
gan.test()
print(" [*] Test finished!")
if args.phase == 'get_results':
gan.GetTest_result()
print(" [*] get your results!")
def main():
# parse arguments
args = parse_args()
if args is None:
exit()
place = fluid.CUDAPlace(0)
# place = fluid.CPUPlace()
with fluid.dygraph.guard(place):
# open session
gan = UGATIT(args)
# build graph
gan.build_model()
if args.phase == 'train':
gan.train()
print(" [*] Training finished!")
if args.phase == 'test':
gan.test()
print(" [*] Test finished!")
def main():
bps.init()
torch.manual_seed(1)
torch.cuda.manual_seed(1)
torch.cuda.set_device(bps.local_rank())
# parse arguments
args = parse_args()
if args is None:
exit()
# open session
gan = UGATIT(args)
# build graph
gan.build_model()
if args.phase == 'train':
gan.train()
print(" [*] Training finished!")
if args.phase == 'test':
gan.test()
print(" [*] Test finished!")
def main():
# parse arguments
args = parse_args()
if args is None:
exit()
# open session
gan = UGATIT(args)
# build graph
gan.build_model()
if args.phase == 'train':
gan.train()
print(" [*] Training finished!")
plt.plot(gan.logit_list_real, label="real_logit", alpha=0.3)
plt.plot(gan.logit_list_fake, label="fake_logit", alpha=0.3)
plt.legend()
plt.show()
if args.phase == 'test':
gan.test(show_real=args.show_real)
print(" [*] Test finished!")