app.py

import gradio as gr
from PIL import Image
import cv2
import numpy as np

#demo_inf_lama = gr.Interface.load(name="spaces/CVPR/lama-example")  #gr.Blocks.load
#img = gr.Interface.load("spaces/multimodalart/latentdiffusion")(ent[0],'50','256','256','1',10)[0] 

#assert demo_inf("Foo", "bar") == "Foo bar"

def get_lama(img1, img2):
  print("********Inside LAMA ********")
  #Img1 = cv2.imread(img1) #Image.open(img1)
  #Img2 = cv2.imread(img2) #Image.open(img2)
  #img1_pil = Image.fromarray(Img1)
  #img2_pil = Image.fromarray(Img2)
  
  #img, mask = gr.Interface.load(name="spaces/CVPR/lama-example")(img1, img2, "automatic (U2net)")
  img, mask = gr.Interface.load(name="spaces/CVPR/lama-example")(img1, img2, "manual")
  
  #img = cv2.imread(img)
  #im.save("./data/data_mask.png")
  #im.save(img)
  im = Image.open(img)
  im.save("./transformed_image.jpg")
  #print("Going back to Workflow")
  #return img, mask
  return im, mask
  return "./transformed_image.jpg", mask
  
#inputs = [gr.inputs.Image(type='pil', label="Original Image"),gr.inputs.Image(type='pil',source="canvas", label="Mask",invert_colors=True),gr.inputs.Radio(choices=["automatic (U2net)","manual"], type="value", default="manual", label="Masking option")]
#outputs = [gr.outputs.Image(type="file",label="output"),gr.outputs.Image(type="pil",label="Mask")]
#io5a = gr.Interface(lambda x: x, gr.Image(source='upload', tool='color-sketch'), gr.Image())

demo = gr.Blocks()

with demo:
  gr.Markdown("<h1><center>Testing</center></h1>")
  gr.Markdown(
        """Testing Inference for Gradio. Work in Progress."""
        )
  with gr.Row():
    in_image = gr.Image(type='filepath')
    #in_image_mask = gr.Image(type='filepath') 
    in_image_mask2 = gr.inputs.Image(type='filepath',source="canvas", label="Mask") #,invert_colors=True) #(type='filepath')
    out_image = gr.outputs.Image(type='pil') #(type='file')
    out_image2 = gr.outputs.Image(type="pil",label="Mask")
    
    b1 = gr.Button("Image Button") 
    #b2 = gr.Button("Text Button") 

    b1.click(get_lama,inputs=[in_image, in_image_mask2], outputs=[out_image,out_image2]) 
    
demo.launch(enable_queue=True, debug=True)