update gradio

app.py

@@ -1,146 +1,81 @@
+import spaces
+import torch
+from diffusers import StableDiffusion3InstructPix2PixPipeline, SD3Transformer2DModel
 import gradio as gr
+import PIL.Image
 import numpy as np
-import random
-from diffusers import DiffusionPipeline
-import torch
+from PIL import Image, ImageOps
+
+
+pipe = StableDiffusion3InstructPix2PixPipeline.from_pretrained("BleachNick/SD3_UltraEdit_w_mask", torch_dtype=torch.float16)
+
+pipe = pipe.to("cuda")
+
+
+
+@spaces.GPU(duration=20)
+def generate(image_mask, prompt, num_inference_steps=50, image_guidance_scale=1.6, guidance_scale=7.5, seed=255):
+    def is_blank_mask(mask_img):
+        # Convert the mask to a numpy array and check if all values are 0 (black/transparent)
+        mask_array = np.array(mask_img.convert('L'))  # Convert to luminance to simplify the check
+        return np.all(mask_array == 0)
+    # Set the seed for reproducibility
+    seed = int(seed)
+    generator = torch.manual_seed(seed)
 
-device = "cuda" if torch.cuda.is_available() else "cpu"
+    img = image_mask["background"].convert("RGB")
+    mask_img = image_mask["layers"][0].getchannel('A').convert("RGB")
 
-if torch.cuda.is_available():
-    torch.cuda.max_memory_allocated(device=device)
-    pipe = DiffusionPipeline.from_pretrained("stabilityai/sdxl-turbo", torch_dtype=torch.float16, variant="fp16", use_safetensors=True)
-    pipe.enable_xformers_memory_efficient_attention()
-    pipe = pipe.to(device)
-else: 
-    pipe = DiffusionPipeline.from_pretrained("stabilityai/sdxl-turbo", use_safetensors=True)
-    pipe = pipe.to(device)
+    # Central crop to desired size
+    desired_size = (512, 512)
 
-MAX_SEED = np.iinfo(np.int32).max
-MAX_IMAGE_SIZE = 1024
+    img = ImageOps.fit(img, desired_size, method=Image.LANCZOS, centering=(0.5, 0.5))
+    mask_img = ImageOps.fit(mask_img, desired_size, method=Image.LANCZOS, centering=(0.5, 0.5))
 
-def infer(prompt, negative_prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps):
+    if is_blank_mask(mask_img):
+        # Create a mask of the same size with all values set to 255 (white)
+        mask_img = PIL.Image.new('RGB', img.size, color=(255, 255, 255))
+    mask_img = mask_img.convert('RGB')
 
-    if randomize_seed:
-        seed = random.randint(0, MAX_SEED)
-        
-    generator = torch.Generator().manual_seed(seed)
-    
     image = pipe(
-        prompt = prompt, 
-        negative_prompt = negative_prompt,
-        guidance_scale = guidance_scale, 
-        num_inference_steps = num_inference_steps, 
-        width = width, 
-        height = height,
-        generator = generator
-    ).images[0] 
-    
-    return image
-
-examples = [
-    "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k",
-    "An astronaut riding a green horse",
-    "A delicious ceviche cheesecake slice",
-]
-
-css="""
-#col-container {
-    margin: 0 auto;
-    max-width: 520px;
-}
+        prompt,
+        image=img,
+        mask_img=mask_img,
+        num_inference_steps=num_inference_steps,
+        image_guidance_scale=image_guidance_scale,
+        guidance_scale=guidance_scale,
+        generator=generator
+    ).images[0]
+
+    return image,mask_img
+
+
+# image_mask_input = gr.ImageMask(label="Input Image", type="pil", brush_color="#000000", elem_id="inputmask",
+#                                 shape=(512, 512))
+image_mask_input = gr.ImageMask(sources='upload',type="pil",label="Input Image: Mask with pen or leave unmasked",transforms=(),layers=False)
+prompt_input = gr.Textbox(label="Prompt")
+num_inference_steps_input = gr.Slider(minimum=0, maximum=100, value=50, label="Number of Inference Steps")
+image_guidance_scale_input = gr.Slider(minimum=0.0, maximum=2.5, value=1.5, label="Image Guidance Scale")
+guidance_scale_input = gr.Slider(minimum=0.0, maximum=17.5, value=12.5, label="Guidance Scale")
+seed_input = gr.Textbox(value="255", label="Random Seed")
+
+inputs = [image_mask_input, prompt_input, num_inference_steps_input, image_guidance_scale_input, guidance_scale_input,
+          seed_input]
+outputs = gr.Image(label="Generated Image")
+
+
+# Custom HTML content
+article_html = """
+<h2>Welcome to the Image Generation Interface</h2>
+<p>This interface allows you to generate images based on a given mask and prompt. Use the sliders to adjust the inference steps and guidance scales, and provide a seed for reproducibility.</p>
 """
 
-if torch.cuda.is_available():
-    power_device = "GPU"
-else:
-    power_device = "CPU"
-
-with gr.Blocks(css=css) as demo:
-    
-    with gr.Column(elem_id="col-container"):
-        gr.Markdown(f"""
-        # Text-to-Image Gradio Template
-        Currently running on {power_device}.
-        """)
-        
-        with gr.Row():
-            
-            prompt = gr.Text(
-                label="Prompt",
-                show_label=False,
-                max_lines=1,
-                placeholder="Enter your prompt",
-                container=False,
-            )
-            
-            run_button = gr.Button("Run", scale=0)
-        
-        result = gr.Image(label="Result", show_label=False)
-
-        with gr.Accordion("Advanced Settings", open=False):
-            
-            negative_prompt = gr.Text(
-                label="Negative prompt",
-                max_lines=1,
-                placeholder="Enter a negative prompt",
-                visible=False,
-            )
-            
-            seed = gr.Slider(
-                label="Seed",
-                minimum=0,
-                maximum=MAX_SEED,
-                step=1,
-                value=0,
-            )
-            
-            randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
-            
-            with gr.Row():
-                
-                width = gr.Slider(
-                    label="Width",
-                    minimum=256,
-                    maximum=MAX_IMAGE_SIZE,
-                    step=32,
-                    value=512,
-                )
-                
-                height = gr.Slider(
-                    label="Height",
-                    minimum=256,
-                    maximum=MAX_IMAGE_SIZE,
-                    step=32,
-                    value=512,
-                )
-            
-            with gr.Row():
-                
-                guidance_scale = gr.Slider(
-                    label="Guidance scale",
-                    minimum=0.0,
-                    maximum=10.0,
-                    step=0.1,
-                    value=0.0,
-                )
-                
-                num_inference_steps = gr.Slider(
-                    label="Number of inference steps",
-                    minimum=1,
-                    maximum=12,
-                    step=1,
-                    value=2,
-                )
-        
-        gr.Examples(
-            examples = examples,
-            inputs = [prompt]
-        )
-
-    run_button.click(
-        fn = infer,
-        inputs = [prompt, negative_prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps],
-        outputs = [result]
-    )
-
-demo.queue().launch()
+demo = gr.Interface(
+    fn=generate,
+    inputs=inputs,
+    outputs=outputs,
+    article=article_html  # Add article parameter
+)
+
+demo.queue().launch()
+