first commit

app.py

@@ -0,0 +1,99 @@
+import gradio as gr
+import numpy as np
+import random
+import torch
+
+from diffusers import StableDiffusionXLPipeline, AutoencoderKL
+from utils import randomize_seed_fn
+
+MAX_SEED = np.iinfo(np.int32).max
+
+def model_load():
+    vae = AutoencoderKL.from_pretrained("madebyollin/sdxl-vae-fp16-fix", torch_dtype=torch.float16)
+    pipe = StableDiffusionXLPipeline.from_pretrained(
+        "stabilityai/stable-diffusion-xl-base-1.0", vae=vae, torch_dtype=torch.float16
+    )
+    # load lora weight
+    pipe.load_lora_weights("jjuun/vivid_color_style")
+
+    return pipe.to('cuda')
+
+
+def sdxl_process(seed, prompt, additional_prompt, negative_prompt, num_steps, guidance_scale):
+    pipe = model_load()
+    generator = torch.Generator("cuda")
+    generator.manual_seed(seed)
+    
+    special_prompt = 'jjj, scratch art style'
+    prompt = f'{special_prompt}, {prompt}, with a black background'
+    output = pipe(prompt, additional_prompt, negative_prompt=negative_prompt, num_inference_steps=num_steps, guidance_scale=guidance_scale,
+                generator=generator).images[0]
+    
+    return output
+
+
+title = "🌈 Colorful illustration" 
+description_en = "🚀 How to use: please make sure to include 'a colorful' in prompt and click Run button!"
+
+
+def create_demo():
+
+    with gr.Blocks() as demo:
+        gr.Markdown(f"<h1 style='text-align: center;'>{title}</h1>")
+        gr.Markdown(f"<h3 style='text-align: center'>{description_en}</h3>")
+
+        with gr.Row():
+            with gr.Column():
+                prompt = gr.Textbox(label="Prompt")
+                run_button = gr.Button("Run")
+                with gr.Accordion("Advanced options", open=False):
+                    
+                    num_steps = gr.Slider(label="Number of steps", minimum=1, maximum=100, value=20, step=1)
+                    guidance_scale = gr.Slider(label="Guidance scale", minimum=0.1, maximum=30.0, value=9.0, step=0.1)
+                    seed = gr.Slider(label="Seed", minimum=0, maximum=MAX_SEED, step=1, value=0)
+                    randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
+                    a_prompt = gr.Textbox(label="Additional prompt", value="")
+                    n_prompt = gr.Textbox(
+                        label="Negative prompt",
+                        value="longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality",
+                    )
+            with gr.Column():
+                result = gr.Image(label="Output")
+                result_seed = gr.Textbox(label="Used seed")
+        
+        gr.Examples(
+            
+            examples= [["a colorful lion", "20", "9", "0", "", "longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality", "examples/lion.png"], 
+                       ["a colorful messi", "20", "9", "0", "", "longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality", "examples/messi.png"]],
+            inputs = [prompt, num_steps, guidance_scale, seed, a_prompt, n_prompt, result]
+        )
+
+        inputs = [
+            seed,
+            prompt,
+            a_prompt,
+            n_prompt,
+            num_steps,
+            guidance_scale,
+        ]
+
+        run_button.click(
+            fn=randomize_seed_fn,
+            inputs=[seed, randomize_seed],
+            outputs=result_seed,
+            queue=False,
+            api_name=False,
+        ).then(
+            fn=sdxl_process,
+            inputs=inputs,
+            outputs=result,
+            api_name=False,
+        )
+    
+
+    return demo
+
+
+if __name__ == "__main__":
+    demo = create_demo()
+    demo.queue().launch()

utils.py

@@ -0,0 +1,23 @@
+import numpy as np
+import random
+import torch
+import os
+
+
+def seed_everything(seed: int = 42):
+    random.seed(seed)
+    np.random.seed(seed)
+    os.environ["PYTHONHASHSEED"] = str(seed)
+    torch.manual_seed(seed)
+    torch.cuda.manual_seed(seed)  # type: ignore
+    torch.backends.cudnn.deterministic = True  # type: ignore
+    torch.backends.cudnn.benchmark = True  # type: ignore
+
+
+def randomize_seed_fn(seed: int, randomize_seed: bool) -> int:
+    MAX_SEED = np.iinfo(np.int32).max
+    
+    if randomize_seed:
+        seed = random.randint(0, MAX_SEED)
+    seed_everything(seed)
+    return seed