Update app.py

app.py

@@ -12,8 +12,67 @@ from PIL import Image
 import gradio as gr
 import os
 
-#Load Magenta Arbitrary Image Stylization network
-hub_module = hub.load('https://tfhub.dev/google/magenta/arbitrary-image-stylization-v1-256/1')
+from helper_functions import *
+
+def style_transfer(input_image, artist):
+
+    style_path_van_gogh = keras.utils.get_file('Starry-Night-canvas-Vincent-van-Gogh-New-1889.jpg', 
+                                           'https://cdn.britannica.com/78/43678-050-F4DC8D93/Starry-Night-canvas-Vincent-van-Gogh-New-1889.jpg')
+    style_path_davinci = keras.utils.get_file('Leonardo_da_Vinci_-_Mona_Lisa_%28La_Gioconda%29_-_WGA12711.jpg', 
+                                          'https://upload.wikimedia.org/wikipedia/commons/f/f2/Leonardo_da_Vinci_-_Mona_Lisa_%28La_Gioconda%29_-_WGA12711.jpg')
+    style_path_dali = keras.utils.get_file('The_Persistence_of_Memory.jpg', 
+                                       'https://upload.wikimedia.org/wikipedia/en/d/dd/The_Persistence_of_Memory.jpg')
+    style_path_monet = keras.utils.get_file('Claude_Monet_-_Water_Lilies_-_Google_Art_Project_%28462013%29.jpg',
+                                        'https://upload.wikimedia.org/wikipedia/commons/a/af/Claude_Monet_-_Water_Lilies_-_Google_Art_Project_%28462013%29.jpg')
+    style_path_picasso = keras.utils.get_file('Picasso_The_Weeping_Woman_Tate_identifier_T05010_10.jpg', 
+                                          'https://upload.wikimedia.org/wikipedia/en/1/14/Picasso_The_Weeping_Woman_Tate_identifier_T05010_10.jpg')
+    style_path_rembrandt = keras.utils.get_file('1259px-The_Nightwatch_by_Rembrandt_-_Rijksmuseum.jpg', 
+                                            'https://upload.wikimedia.org/wikipedia/commons/thumb/9/94/The_Nightwatch_by_Rembrandt_-_Rijksmuseum.jpg/1259px-The_Nightwatch_by_Rembrandt_-_Rijksmuseum.jpg')
+    
+    #set dimensions of input image                                        
+    oc_max_dim = 1080
+
+    #set parameters for each choice of artist
+    if artist == "Vincent van Gogh":
+        style_max_dim = 442
+        style_path = style_path_van_gogh
+    elif artist == "Claude Monet":
+        style_max_dim = 256
+        style_path = style_path_monet
+    elif artist == "Leonardo da Vinci":
+        style_max_dim = 442
+        style_path = style_path_davinci
+    elif artist == "Rembrandt":
+        style_max_dim = 256
+        style_path = style_path_rembrandt
+    elif artist == "Pablo Picasso":
+        style_max_dim = 256
+        style_path = style_path_picasso
+    elif artist == "Salvador Dali":
+        style_max_dim = 512
+        style_path = style_path_dali
+    
+    #load content and style images    
+    content_image = load_img(input_image, max_dim=oc_max_dim)
+    style_image = load_img(style_path, style_max_dim)
+    
+    #Load Magenta Arbitrary Image Stylization network
+    hub_module = hub.load('https://tfhub.dev/google/magenta/arbitrary-image-stylization-v1-256/1')
+    
+    #Pass content and style images as arguments in TensorFlow Constant object format
+    stylized_image = hub_module(tf.constant(content_image), tf.constant(style_image))[0]
+    
+    return stylized_image
+    
+app = gr.Interface(
+    style_transfer,
+    [gr.Image(type='pil'), gr.Radio(["Vincent van Gogh", "Claude Monet", "Leonardo da Vinci", "Rembrandt", "Pablo Picasso", "Salvador Dali"])],
+    gr.Image(type='pil'),
+    title="Artist Style Transfer Tool",
+    description="Make your own art in the style of six famous artists using pretrained neural networks and deep learning!"
+    #article="https://arxiv.org/abs/1705.06830"
+)
+    
 """
 def inference(input_image):
     preprocess = transforms.Compose([
@@ -63,14 +122,14 @@ def inference(input_image):
 ##title = "Artist Classifier"
 ##description = "Gradio demo for MOBILENET V2, Efficient networks optimized for speed and memory, with residual blocks. To use it, simply upload your image, or click one of the examples to load them. Read more at the links below."
 ##article = "<p style='text-align: center'><a href='https://arxiv.org/abs/1801.04381'>MobileNetV2: Inverted Residuals and Linear Bottlenecks</a> | <a href='https://github.com/pytorch/vision/blob/master/torchvision/models/mobilenet.py'>Github Repo</a></p>"
-
+"""
 def greet(name):
     return "Hello " + name + "!!"
 
 demo = gr.Interface(fn=greet, inputs="text", outputs="text")
 
 demo.launch()
-
+"""
 #examples = [
 #            ['dog.jpg']
 #]