Update app.py

app.py

@@ -1,11 +1,13 @@
 import spaces
-from diffusers import ControlNetModel, StableDiffusionXLControlNetPipeline, EulerAncestralDiscreteScheduler
-import torch
-import gradio as gr
+from diffusers import ControlNetModel, StableDiffusionXLControlNetPipeline, AutoencoderKL, EulerAncestralDiscreteScheduler
+from diffusers.utils import load_image
 from PIL import Image
+import torch
 import numpy as np
+import cv2
+import gradio as gr
+from torchvision import transforms 
 
-# Load models
 controlnet = ControlNetModel.from_pretrained(
     "briaai/BRIA-2.2-ControlNet-Recoloring",
     torch_dtype=torch.float16
@@ -19,7 +21,6 @@ pipe = StableDiffusionXLControlNetPipeline.from_pretrained(
     low_cpu_mem_usage=True,
     offload_state_dict=True,
 ).to('cuda')
-
 pipe.scheduler = EulerAncestralDiscreteScheduler(
     beta_start=0.00085,
     beta_end=0.012,
@@ -27,52 +28,67 @@ pipe.scheduler = EulerAncestralDiscreteScheduler(
     num_train_timesteps=1000,
     steps_offset=1
 )
-
+# pipe.enable_freeu(b1=1.1, b2=1.1, s1=0.5, s2=0.7)
+# pipe.enable_xformers_memory_efficient_attention()
 pipe.force_zeros_for_empty_prompt = False
 
 def resize_image(image):
     image = image.convert('RGB')
     current_size = image.size
-    transform = gr.Image(height=1024, width=1024, keep_aspect_ratio=True, source="upload", tool="editor")
-    resized_image = transform.postprocess(image)
+    if current_size[0] > current_size[1]:
+        center_cropped_image = transforms.functional.center_crop(image, (current_size[1], current_size[1]))
+    else:
+        center_cropped_image = transforms.functional.center_crop(image, (current_size[0], current_size[0]))
+    resized_image = transforms.functional.resize(center_cropped_image, (1024, 1024))
     return resized_image
 
-@spaces.GPU(enable_queue=True)
-def generate_image(input_image, prompt, controlnet_conditioning_scale):
-    # Always use a random seed for diversity in outputs
-    seed = np.random.randint(2147483647)
-    generator = torch.Generator("cuda").manual_seed(seed)  
-    
-    # Resize and prepare the image
-    input_image = resize_image(input_image)
-    grayscale_image = input_image.convert('L').convert('RGB')
-    
-    # Generate the image with fixed 30 steps
+
+@spaces.GPU
+def generate_(prompt, negative_prompt, grayscale_image, num_steps, controlnet_conditioning_scale, seed):
+    generator = torch.Generator("cuda").manual_seed(seed)    
     images = pipe(
-        prompt=prompt, 
-        image=grayscale_image, 
-        num_inference_steps=30, 
-        controlnet_conditioning_scale=float(controlnet_conditioning_scale),
-        generator=generator,
+    prompt, negative_prompt=negative_prompt, image=grayscale_image, num_inference_steps=num_steps, controlnet_conditioning_scale=float(controlnet_conditioning_scale),
+    generator=generator,
     ).images
+    return images
+
+@spaces.GPU
+def process(input_image, prompt, negative_prompt, num_steps, controlnet_conditioning_scale, seed):
+    
+    # resize input_image to 1024x1024
+    input_image = resize_image(input_image)
     
-    return images[0]
+    grayscale_image = input_image.convert('L').convert('RGB')
+  
+    images = generate_(prompt, negative_prompt, grayscale_image, num_steps, controlnet_conditioning_scale, seed)
 
-# Gradio 
-description = "Anything to Anything. Transform anything to anything with just a prompt!"
+    return [grayscale_image,images[0]]
+    
+block = gr.Blocks().queue()
 
-with gr.Blocks() as demo:
-    gr.Markdown("<h1><center>Image Transformation with Bria Recolor ControlNet</center></h1>")
-    gr.Markdown(description)
+with block:
+    gr.Markdown("## BRIA 2.2 ControlNet Recoloring")
+    gr.HTML('''
+      <p style="margin-bottom: 10px; font-size: 94%">
+        This is a demo for ControlNet Recoloring that using
+        <a href="https://huggingface.co/briaai/BRIA-2.2" target="_blank">BRIA 2.2 text-to-image model</a> as backbone. 
+        Trained on licensed data, BRIA 2.2 provide full legal liability coverage for copyright and privacy infringement.
+      </p>
+    ''')
     with gr.Row():
         with gr.Column():
-            input_image = gr.Image(label='Upload your image', type="pil")
-            prompt = gr.Textbox(label='Enter your prompt')
+            input_image = gr.Image(sources=None, type="pil") # None for upload, ctrl+v and webcam
+            prompt = gr.Textbox(label="Prompt")
+            negative_prompt = gr.Textbox(label="Negative prompt", value="Logo,Watermark,Text,Ugly,Morbid,Extra fingers,Poorly drawn hands,Mutation,Blurry,Extra limbs,Gross proportions,Missing arms,Mutated hands,Long neck,Duplicate,Mutilated,Mutilated hands,Poorly drawn face,Deformed,Bad anatomy,Cloned face,Malformed limbs,Missing legs,Too many fingers")
+            num_steps = gr.Slider(label="Number of steps", minimum=25, maximum=100, value=50, step=1)
             controlnet_conditioning_scale = gr.Slider(label="ControlNet conditioning scale", minimum=0.1, maximum=2.0, value=1.0, step=0.05)
-            submit_button = gr.Button('Transform Image')
+            seed = gr.Slider(label="Seed", minimum=0, maximum=2147483647, step=1, randomize=True,)
+            run_button = gr.Button(value="Run")
+            
+            
         with gr.Column():
-            output_image = gr.Image(label='Transformed Image')
-
-    submit_button.click(fn=generate_image, inputs=[input_image, prompt, controlnet_conditioning_scale], outputs=output_image)
+            result_gallery = gr.Gallery(label='Output', show_label=False, elem_id="gallery", columns=[2], height='auto')
+    ips = [input_image, prompt, negative_prompt, num_steps, controlnet_conditioning_scale, seed]
+    run_button.click(fn=process, inputs=ips, outputs=[result_gallery])
 
-demo.queue().launch()
+block.launch(debug = True)