Update app.py

app.py

@@ -1,6 +1,4 @@
 import os
-os.system("conda install pytorch torchvision -c pytorch")
-os.system("pip install torch")
 import gradio as gr
 from PIL import Image
 import torch
@@ -16,24 +14,20 @@ def inference(img):
     img = img.resize((basewidth,hsize), Image.ANTIALIAS)
     img.save("test/1.jpg", "JPEG")
     os.system('python main_test_swinir.py --task real_sr --model_path experiments/pretrained_models/003_realSR_BSRGAN_DFO_s64w8_SwinIR-M_x4_GAN.pth --folder_lq test --scale 4')
-    
-    # Load the output image file as an Image object
-    output_img = Image.open('results/swinir_real_sr_x4/1_SwinIR.png')
-    
-    # Return the output image
-    return output_img
+    return 'results/swinir_real_sr_x4/1_SwinIR.png'
         
 title = "SwinIR"
 description = "Gradio demo for SwinIR. SwinIR achieves state-of-the-art performance on six tasks: image super-resolution (including classical, lightweight and real-world image super-resolution), image denoising (including grayscale and color image denoising) and JPEG compression artifact reduction. See the paper and project page for detailed results below. Here, we provide a demo for real-world image SR.To use it, simply upload your image, or click one of the examples to load them."
 article = "<p style='text-align: center'><a href='https://arxiv.org/abs/2108.10257' target='_blank'>SwinIR: Image Restoration Using Swin Transformer</a> | <a href='https://github.com/JingyunLiang/SwinIR' target='_blank'>Github Repo</a></p>"
 
-
+examples=[['ETH_LR.png']]
 gr.Interface(
     inference, 
     [gr.inputs.Image(type="pil", label="Input")], 
-    gr.outputs.Image(type="pil", label="Output"),
+    gr.outputs.Image(type="file", label="Output"),
     title=title,
     description=description,
     article=article,
-    enable_queue=True
-    ).launch(debug=True)
+    enable_queue=True,
+    examples=examples
+    ).launch(debug=True)