Deep daze is a command line tool that uses openai clip and siren to generate images from text. It can generate corresponding images by using simple language to describe image content.
![shattered plates on the grass
(broken dishes on the grass) ](https://gitee.com/kaixiaoyan/…
Download and install
Deep daze is a python command-line tool, so you need to install Python in the environment you use, and then execute the following command to install it:
$ pip install deep-daze
Easy to use
The use of deep daze is also very simple. Just remember an image command, such as:
$ imagine "a house in the forest"
In windows, you need to use the administrator to open the CMD window.
If the memory is large enough, you can add the — Deep option to get higher quality pictures
$ imagine "shattered plates on the ground" --deeper
Deep daze has the following options:
--img=IMAGE_PATH Default: none. To optimize the path of PNG / JPG image or PIL image. --encoding=ENCODING Default: none. User created custom clip encoding. If used, replaces any text or image used. --create_story=CREATE_STORY Default value: false. If you enable this feature, you can use text longer than 77 characters to create picture stories. --story_start_words=STORY_START_WORDS Default value: 5. Only in create_ Used when story is true. --story_words_per_epoch=STORY_WORDS_PER_EPOCH Default value: 5. Only in create_ Used when story is true. --story_separator： Default: None Only in create_ Used when story is true.定义一个类似.的分隔符。 --lower_bound_cutout=LOWER_BOUND_CUTOUT Default: 0.1 The sampling lower limit of the size of random incision of each batch of siren images. Should be less than 0 . 8。 --upper_bound_cutout=UPPER_BOUND_CUTOUT Default: 1.0 The sampling upper limit of the size of random incision of each batch of siren images. It should be kept at 1.0. --saturate_bound=SATURATE_BOUND Default: false If true, lower is set during training_ BOUND_ The cutoff increased linearly to 0.75. --learning_rate=LEARNING_RATE Default value: 1e-05 Learning rate of neural network. --num_layers=NUM_LAYERS Default value: 16 The number of hidden layers of siren neural network. --batch_size=BATCH_SIZE Default value: 4 The number of images passed to Siren before the loss is calculated. Reducing this value may reduce memory and accuracy. --gradient_accumulate_every=GRADIENT_ACCUMULATE_EVERY Default value: 4 The weighted loss of n samples is calculated. Increasing this value helps to improve accuracy with a smaller batch size. --epochs=EPOCHS Default value: 20 The number of times to run. -- iterations = iteration Default value: 1050 The number of times the sum and backpropagation losses are calculated in a given period. --save_every=SAVE_EVERY Default value: 100 Each iteration of the generated image is a multiple of this number. --image_width = IMAGE_WIDTH Default value: 512 The required image resolution. --deeper=DEEPER Default: false The siren neural network with 32 hidden layers is used. --overwrite=OVERWRITE Default: false Whether to overlay the existing generated image with the same name. --save_progress=SAVE_PROGRESS Default: false Whether to save the image generated before siren training. --seed=SEED Type: optional Default: None The seed to be used is for deterministic operation. --open_folder=OPEN_FOLDER Default value: true Whether to open the folder for the generated images. --save_date_time=SAVE_DATE_TIME Default: false Save the file with a timestamp. For example, '% Y% m% d -% H% m% s-my'_ phrase_ here` --start_image_path= TART_IMAGE_PATH Default: None First train students on the original image, then turn to text input --start_image_train_iters=START_IMAGE_TRAIN_ITERS Default value: 50 The number of initial training on the initial image --theta_initial=THETA_INITIAL Default value: 30.0 Parameters describing the frequency of color space. Only for the first layer of the network. --theta_hidden = THETA_INITIAL Default value: 30.0 Parameters describing the frequency of color space. Only for the hidden layer of the network. --save_gif = SAVE_GIF Default: false Whether to save GIF animation for the build process. Only in save_ Valid when progress is set to true.
- Training synthesis based on one picture
$ imagine 'a clear night sky filled with stars' --start_image_path ./cloudy-night-sky.jpg
- Calling with Python
from deep_daze import Imagine imagine = Imagine( text = 'cosmic love and attention', num_layers = 24, ) imagine()
- Save every four iterations. Save the image in this format: insert\_ text\_ here.00001.png，insert\_ text\_ here.00002.png，…
imagine = Imagine( text=text, save_every=4, save_progress=True )
- Create a file with a timestamp and serial number
imagine = Imagine( text=text, save_every=4, save_progress=True, save_date_time=True, )
Open source outpostShare popular, interesting and practical open source projects everyday. Participate in the maintenance of 100000 + star’s open source technology resource library, including python, Java, C / C + +, go, JS, CSS Node.js , PHP,. Net, etc.