Update app.py

app.py

@@ -1,70 +1,89 @@
 import gradio as gr
 import numpy as np
+import PIL.Image
+from PIL import Image
 import random
-from diffusers import DiffusionPipeline
+from diffusers import ControlNetModel, StableDiffusionXLControlNetPipeline, StableDiffusionXLPipeline, AutoencoderKL
+from diffusers import DDIMScheduler, EulerAncestralDiscreteScheduler
+#from diffusers.utils import load_image
+
+import cv2
 import torch
+import spaces
+
+
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+
+controlnet = ControlNetModel.from_pretrained(
+    #"2vXpSwA7/test_controlnet2/CN-anytest_v4-marged_am_dim256.safetensors",
+    "xinsir/controlnet-scribble-sdxl-1.0",
+    torch_dtype=torch.float16
+    #from_tf=False,  
+    #variant="safetensors"
+)
 
-device = "cuda" if torch.cuda.is_available() else "cpu"
+vae = AutoencoderKL.from_pretrained("madebyollin/sdxl-vae-fp16-fix", torch_dtype=torch.float16)
 
-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)
+pipe = StableDiffusionXLControlNetPipeline.from_pretrained(
+    "yodayo-ai/holodayo-xl-2.1",
+    controlnet=controlnet,
+    vae=vae,
+    torch_dtype=torch.float16,
+)
+
+pipe.scheduler = EulerAncestralDiscreteScheduler.from_config(pipe.scheduler.config)
+pipe.to(device)
 
 MAX_SEED = np.iinfo(np.int32).max
-MAX_IMAGE_SIZE = 1024
+MAX_IMAGE_SIZE = 1216
 
-def infer(prompt, negative_prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps):
+    
+@spaces.GPU
+def infer(prompt, negative_prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps, image: PIL.Image.Image) -> PIL.Image.Image:
+    width, height = image['composite'].size
+    ratio = np.sqrt(1024. * 1024. / (width * height))
+    new_width, new_height = int(width * ratio), int(height * ratio)
+    image = image['composite'].resize((new_width, new_height))
+    print(image)
 
     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",
-]
+    output_image = pipe(
+        prompt=prompt + ", masterpiece, best quality, very aesthetic, absurdres",
+        negative_prompt=negative_prompt,
+        image=image,
+        controlnet_conditioning_scale=1.0,
+        guidance_scale=guidance_scale,
+        num_inference_steps=num_inference_steps,
+        width=new_width,
+        height=new_height,
+        generator=generator
+    ).images[0]
+
+    return output_image
+
 
-css="""
+css = """
 #col-container {
     margin: 0 auto;
     max-width: 520px;
 }
 """
 
-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}.
+        gr.Markdown("""
+        # Text-to-Image Demo
+        using : 
+        [Holodayo XL 2.1](https://huggingface.co/yodayo-ai/holodayo-xl-2.1), 
+        [scribble](https://huggingface.co/xinsir/controlnet-scribble-sdxl-1.0)
         """)
-        
+
         with gr.Row():
-            
             prompt = gr.Text(
                 label="Prompt",
                 show_label=False,
@@ -72,20 +91,22 @@ with gr.Blocks(css=css) as demo:
                 placeholder="Enter your prompt",
                 container=False,
             )
-            
+
             run_button = gr.Button("Run", scale=0)
-        
+
+        image = gr.ImageEditor(type="pil", image_mode="L", crop_size=(512, 512))
         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,
+                value="nsfw, (low quality, worst quality:1.2), very displeasing, 3d, watermark, signature, ugly, poorly drawn"
             )
-            
+
             seed = gr.Slider(
                 label="Seed",
                 minimum=0,
@@ -93,54 +114,47 @@ with gr.Blocks(css=css) as demo:
                 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,
+                    value=1024,#832,
                 )
-                
+
                 height = gr.Slider(
                     label="Height",
                     minimum=256,
                     maximum=MAX_IMAGE_SIZE,
                     step=32,
-                    value=512,
+                    value=1024,#1216,
                 )
-            
+
             with gr.Row():
-                
                 guidance_scale = gr.Slider(
                     label="Guidance scale",
                     minimum=0.0,
-                    maximum=10.0,
+                    maximum=20.0,
                     step=0.1,
-                    value=0.0,
+                    value=7,
                 )
-                
+
                 num_inference_steps = gr.Slider(
                     label="Number of inference steps",
                     minimum=1,
-                    maximum=12,
+                    maximum=28,
                     step=1,
-                    value=2,
+                    value=28,
                 )
-        
-        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]
+
+    run_button.click(#lambda x: None, inputs=None, outputs=result).then(
+        fn=infer,
+        inputs=[use_image, prompt, negative_prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps,image],
+        outputs=[result]
     )
 
-demo.queue().launch()
+demo.queue().launch()