app.py

import gradio as gr
import numpy as np
import matplotlib.pyplot as plt
from tensorflow.keras.datasets import mnist
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense, Reshape, Flatten
from tensorflow.keras.layers import Conv2D, Conv2DTranspose
from tensorflow.keras.layers import LeakyReLU, Dropout
from tensorflow.keras.optimizers import Adam

np.random.seed(42)

# Загрузка и предобработка данных MNIST
(X_train, _), (_, _) = mnist.load_data()
X_train = (X_train.astype(np.float32) - 127.5) / 127.5
X_train = np.expand_dims(X_train, axis=3)

# Создание и компиляция модели GAN
generator = Sequential()
generator.add(Dense(7 * 7 * 256, input_dim=100))
generator.add(LeakyReLU(alpha=0.2))
generator.add(Reshape((7, 7, 256)))
generator.add(Conv2DTranspose(128, (5, 5), strides=(1, 1), padding='same'))
generator.add(LeakyReLU(alpha=0.2))
generator.add(Conv2DTranspose(64, (5, 5), strides=(2, 2), padding='same'))
generator.add(LeakyReLU(alpha=0.2))
generator.add(Conv2DTranspose(1, (5, 5), strides=(2, 2), padding='same'))

discriminator = Sequential()
discriminator.add(Conv2D(64, (5, 5), strides=(2, 2), padding='same', input_shape=(28, 28, 1)))
discriminator.add(LeakyReLU(alpha=0.2))
discriminator.add(Dropout(0.3))
discriminator.add(Conv2D(128, (5, 5), strides=(2, 2), padding='same'))
discriminator.add(LeakyReLU(alpha=0.2))
discriminator.add(Dropout(0.3))
discriminator.add(Flatten())
discriminator.add(Dense(1, activation='sigmoid'))

discriminator.compile(loss='binary_crossentropy', optimizer=Adam(learning_rate=0.0002, beta_1=0.5), metrics=['accuracy'])
discriminator.trainable = False
gan = Sequential()
gan.add(generator)
gan.add(discriminator)
gan.compile(loss='binary_crossentropy', optimizer=Adam(learning_rate=0.0002, beta_1=0.5))

# Функция для генерации изображений на основе запроса
def generate_images(prompt):
    num_images = 10
    noise = np.random.normal(0, 1, (num_images, 100))
    generated_images = generator.predict(noise)
    generated_images = (generated_images * 127.5 + 127.5).astype(np.uint8)
    return generated_images.reshape((num_images, 28, 28))

# Адаптация функции под Gradio
def gradio_generate_images(prompt):
    images = generate_images(prompt)
    image_list = []
    for img in images:
        image_list.append(img)
    return image_list

# Запуск Gradio приложения
iface = gr.Interface(
    fn=gradio_generate_images,
    inputs="text",
    outputs="image",
    interpretation="default",
    title="GAN Image Generation Demo",
    description="Enter a prompt and generate images based on the prompt using GAN.",
    example="smiley face"
)
iface.launch()