def my_form_post():
#text = request.form['text']
# imageName = text.split("/")[-1]
# contentImagePath = "images/input/"+imageName
# outputImagePath = "images/output/"+imageName
# # print(file_path)
# if not os.path.exists(outputImagePath):
# f = open(contentImagePath, 'wb')
# f.write(requests.get(text).content)
# f.close()
#style_transfer("images/profile.jpg")
# data = {}
# # try:
# data = request.get_json()['data']
# # except Exception:
# # return jsonify(status_code='400', msg='Bad Request'), 400
# data = base64.b64decode(data)
# image = io.BytesIO(data)
style_transfer('profile.jpg','style.jpg')
result_dir = os.getcwd() + "/static/css/images"
urls = [f for f in os.listdir(result_dir)]
return render_template("show_images.html", urls = urls)
def main():
content_placeholder = tf.placeholder(tf.float32, shape=(1, 256, 256, 3))
generated_images = model.style_transfer(content_placeholder)
output_format = tf.cast(generated_images, tf.uint8)
jpegs = tf.map_fn(lambda image: tf.image.encode_jpeg(image),
output_format,
dtype=tf.string)
with tf.Session() as sess:
file = tf.train.latest_checkpoint(model_path)
if not file:
print('Could not find trained model in %s' % model_path)
return
print('Using model from %s' % file)
saver = tf.train.Saver()
saver.restore(sess, file)
images_t = sess.run(
jpegs, feed_dict={content_placeholder: get_image(CONTENT_IMAGE)})
with open('res.jpg', 'wb') as f:
f.write(images_t[0])
os.mkdir(flags.summary_dir)
if flags.mode == 'train':
with tf.device('/gpu:0'):
list1_ = os.listdir(flags.train_dir1)
list1 = [os.path.join(flags.train_dir1,_) for _ in list1_ if _.split('.')[1]=='jpg']
list2_ = os.listdir(flags.train_dir2)
list2 = [os.path.join(flags.train_dir2, _) for _ in list2_ if _.split('.')[1] == 'jpg']
batch1 = imgs_loder(list1,flags)
batch2 = imgs_loder(list2,flags)
net = style_transfer(batch1,batch2,flags)
tf.summary.scalar('discrim_loss',net.discrim_loss)
tf.summary.scalar('content_loss', net.discrim_loss)
tf.summary.scalar('adversarial_loss', net.adversarial_loss)
tf.summary.scalar('learning_rate', net.learning_rate)
saver = tf.train.Saver(max_to_keep=10)
var_list = tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES)
var_list2 = tf.get_collection(tf.GraphKeys.TRAINABLE_VARIABLES, scope='generator')
weight_initiallizer = tf.train.Saver(var_list2)
decoder.to(device)
T = transforms.Compose([
transforms.ToTensor(),
])
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(message)s')
for content_path in content_paths:
for style_path in style_paths:
logging.info("Transfering from " + style_path + " to " +
content_path + " ...")
content = T(Image.open(content_path).convert('RGB'))
content = content.to(device).unsqueeze(0)
style = T(Image.open(style_path).convert('RGB'))
style = style.to(device).unsqueeze(0)
with torch.no_grad():
output = model.style_transfer(encoder, decoder, content, style,
args.alpha)
output = output.cpu()
output_name = '{:s}/{:s}_stylized_{:s}{:s}'.format(
args.output,
splitext(basename(content_path))[0],
splitext(basename(style_path))[0], ".jpg")
save_image(output, output_name)