def predict():
data = {'Success': False}
if request.files.get('image'):
now = time.strftime("%Y-%m-%d-%H_%M_%S", time.localtime(time.time()))
image = request.files['image'].read()
image = Image.open(io.BytesIO(image))
image = image_transform(InputSize)(image).numpy()
# 将数组以C语言存储顺序存储
image = image.copy(order="C")
# 生成图像ID
k = str(uuid.uuid4())
d = {"id": k, "image": base64_encode_image(image)}
# print(d)
db.rpush(ImageQueue, json.dumps(d))
# 运行服务
while True:
# 获取输出结果
output = db.get(k)
# print(output)
if output is not None:
output = output.decode("utf-8")
data["predictions"] = json.loads(output)
db.delete(k)
break
time.sleep(ClientSleep)
data["success"] = True
return jsonify(data)
def baidu_image2str_url(uuid_url_dict={}, types="characters"):
from aip import AipOcr
client = AipOcr(Baidu_APP_ID, Baidu_API_KEY,
Baidu_SECRET_KEY) # create a connection
options = {}
options["probability"] = "true"
uuid_text_dict = {}
for uuid, url in uuid_url_dict.items():
ret = ""
resp = client.basicGeneralUrl(url, options) # url
# print(resp)
if "error_msg" in resp:
print("url recognition failed! Using local model. url: " + url)
# print(resp)
if resp["error_msg"] == "url response invalid" or resp[
"error_msg"] == "image size error":
#request for the image of url, convert to valid format
image_path = image_transform(url_img_download(url))
print(image_path)
uuid_text_dict[uuid] = baidu_image2str_local(image_path)
else:
uuid_text_dict[uuid] = ""
else:
for tex in resp["words_result"]:
if tex["probability"]["average"] > 0.85:
ret = ret + tex["words"]
# print(ret)
uuid_text_dict[uuid] = ret
return uuid_text_dict
def set_info(self, data):
self.uuid_label.setText(data['uuid'])
if os.path.exists(data['local_path']):
img_path = data['local_path']
else:
img_path = utils.image_transform(
utils.url_img_download(data['src']))
self.set_image(img_path)
if data['resolved']:
self.text_box.setText(data['plain_text'])
else:
self.text_box.setText('')
self.text_box.setFocus()
def get_next_batch_data(self):
if self.getting == False:
self.getting = True
data = self.driver.load_unchecked_img(10, self.size - self.checked)
count = 0
for d in data:
count = count + 1
img_path = utils.image_transform(
utils.url_img_download(d['src']))
d['local_path'] = img_path
if 'plain_text' not in d.keys():
d['plain_text'] = ''
self.data.append(d)
self.size = self.size + count
self.getting = False
e = utils.create_experiment("experiments/dcgan4_celeba")
# Hyperparameters
e.params["shuffle"] = True # Shuffle the contents in the dataset
e.params["num_workers"] = 4 # Number of worker threads for dataloader
e.params["batch_size"] = 128 # Size of one batch during training
e.params["nc"] = 3 # Number of channels in the training images (color RGB uses 3 channels)
e.params["nz"] = 100 # Size of z latent vector (i.e. size of generator input)
e.params["im_size"] = 64 # Size of the images discriminated and generated.
e.params["num_epochs"] = e.input_int("number of epochs", 5) # Number of epochs
e.params["lr"] = 0.0002 # Learning rate for optimizer
e.params["betas"] = (0.5, 0.999) # Betas hyperparameter for Adam optimizers
e.params["patience"] = 7 # Number of epochs to wait before early stopping
# Setup the CIFAR10 dataset
transform = utils.image_transform(e.params["im_size"])
data_dir = "data/celeba/"
# WARNING DOWNLOAD IS 1.4 GB in size!!!
# train_dataset = datasets.CelebA(data_dir, split="train", download=False, transform=transform)
train_dataset = datasets.ImageFolder(data_dir, transform=transform)
e.setup_dataloader((train_dataset, None, None))
# Plot a subset of the training dataset
utils.plot_data_subset(e.fname("dataset_image.png"), train_dataset, show_labels=False)
# Setup the two models
e.generator = models.dcgan4_generator(e)
e.discriminator = models.dcgan4_discriminator(e)
fig, axes = plt.subplots(2, len(original_images))
for i in range(len(original_images)):
axes[0, i].imshow(original_images[i])
axes[1, i].imshow(generated_images[i])
plt.show()
if __name__ == '__main__':
EPOCHS = 2
BATCH_SIZE = 4
ROOT = os.getcwd()
SHOW_EVERY = 500
TRAIN_IMAGES_FOLDER = os.path.join(ROOT, 'train2014')
STYLE_IMAGE_FOLDER = os.path.join(ROOT, 'StyleImages')
content_dataset = MSCOCODataset(TRAIN_IMAGES_FOLDER, image_transform())
style_dataset = StyleImageDataset(STYLE_IMAGE_FOLDER, image_transform())
style_transfer_net = StyleTransferNet()
optimizer = Adam(style_transfer_net.parameters(), lr=1e-3)
C = 100000.0
S = 700000.0
COLOR = 20000.0
DEVICE = 'cuda'
PKL_NAME = 'vangogh_YIQ'
train(style_transfer_net, EPOCHS, BATCH_SIZE, content_dataset,
style_dataset, optimizer, C, S, COLOR, DEVICE, PKL_NAME)