added gfp gan

app.py

@@ -7,15 +7,20 @@ import torchvision.transforms as transforms
 import PIL
 from PIL import Image
 from PIL import ImageFile
+from pathlib import Path
+import shutil
 import math
 import os
 import torch.nn.functional as F
 from rich.panel import Panel
 from rich.columns import Columns
 from rich.console import Console
+from models.gfpgan import gfpgan_predict
 
 os.environ["CUDA_VISIBLE_DEVICES"] = "1"
 device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
+
+
 model1 = my_model(en_feature_num=48,
                      en_inter_num=32,
                      de_feature_num=64,
@@ -33,7 +38,10 @@ def default_toTensor(img):
     composed_transform = transforms.Compose(t_list)
     return composed_transform(img)
 
-def predict1(img):
+def predict(img):
+    
+    
+    
     in_img = transforms.ToTensor()(img).to(device).unsqueeze(0)
     b, c, h, w = in_img.size()
     # pad image such that the resolution is a multiple of 32
@@ -75,7 +83,32 @@ def img_pad(x,  w_pad, h_pad, w_odd_pad, h_odd_pad):
     return y
 
 
-title = "image restore"
+
+
+def predict_gfpgan(img):
+    
+    with Console().status("[red] using [green] GFP-GAN v1.4", spinner="aesthetic"):    
+        # if image already exists with this name then delete it
+        if Path("input_image_gfpgan.jpg").exists():
+            os.remove("input_image_gfpgan.jpg")
+        # save incoming PIL image to disk
+        img.save("input_image_gfpgan.jpg")
+        
+        out = gfpgan_predict(img)
+        Console().print(out)
+        
+        return img
+
+
+
+
+
+
+
+
+
+
+title = "image enhancement"
 description = " The model was trained to remove the moire patterns from images! Max input image resolution is \
 up to 4K resolution\
 <br /> \
@@ -85,6 +118,10 @@ It takes time to perform inference \
 article = "Reference:\n Towards Efficient and Scale-Robust Ultra-High-Definition Image Demoiréing. CVMI Lab, Nov. 21, 2022. Accessed: Nov. 21, 2022. [Online]. Available: https://github.com/CVMI-Lab/UHDM"
 
 
+# ==========================================================================
+#                             example images                                  
+# ==========================================================================
+
 files = [
     '0001_01.jpg',
     '0002_01.jpg',
@@ -676,14 +713,37 @@ files = [os.path.join("n000129", k) for k in files]
 
 Console().print(Columns ([Panel.fit(f"{k}", style="red on black") for k in files]))
 
-iface1 = gr.Interface(fn=predict1,
+
+# # --------------------------------------------------------------------------
+# #                              making interfaces for models                        
+# # --------------------------------------------------------------------------
+
+iface1 = gr.Interface(fn=predict,
                      inputs=gr.Image(type="pil"),
                      outputs=gr.Image(type="pil"),
                      examples=files,
                     title = title,
                     # description = description,
-                    article = article
+                    article = article,
+                    allow_flagging="auto"
                     )
 
-# iface1.queue(concurrency_count=3)
-iface1.launch()
+gfpgan = gr.Interface(fn=predict_gfpgan,
+                     inputs=gr.Image(type="pil"),
+                     outputs=gr.Image(type="pil"),
+                     examples=files,
+                    title = "GFP-GAN v 1.4",
+                    # description = description,
+                    article = "Practical face restoration algorithm for old photos",
+                    allow_flagging="auto"
+                    )
+
+
+
+
+# setting queues for all models
+iface1.queue(concurrency_count=3)
+gfpgan.queue(concurrency_count=3)
+
+# launching the interface in parallel model
+gr.Parallel(iface1, gfpgan).launch()

models/gfpgan.py

@@ -0,0 +1,19 @@
+import replicate
+
+
+model = replicate.models.get("tencentarc/gfpgan")
+version = model.versions.get("9283608cc6b7be6b65a8e44983db012355fde4132009bf99d976b2f0896856a3")
+
+
+def gfpgan_predict(input_image : str):
+    '''
+    takes path to image as input and returns enhanced image
+
+    Parameters
+    ----------
+    input_image : str
+        path to the image file
+    '''    
+    output = version.predict(img=input_image, version='v1.4', scale=1.0)
+    
+    return output

requirements.txt

@@ -1,6 +1,7 @@
-numpy==1.21.5
-torch>=1.9.0
-opencv-python==4.5.5.64
-scikit-image==0.19.2
-torchvision==0.1.8
-
+gradio==3.11.0
+opencv_python_headless==4.6.0.66
+Pillow==9.3.0
+replicate==0.4.0
+rich==12.6.0
+torch==1.11.0
+torchvision==0.12.0