import os
import random
import uuid
import gradio as gr
import numpy as np
from PIL import Image
import spaces
import torch
from diffusers import StableDiffusion3Pipeline, DPMSolverMultistepScheduler, AutoencoderKL, StableDiffusion3Img2ImgPipeline
from huggingface_hub import snapshot_download
huggingface_token = os.getenv("HUGGINGFACE_TOKEN")
model_path = snapshot_download(
repo_id="stabilityai/stable-diffusion-3-medium",
revision="refs/pr/26",
repo_type="model",
ignore_patterns=["*.md", "*..gitattributes"],
local_dir="stable-diffusion-3-medium",
token=huggingface_token, # type a new token-id.
)
DESCRIPTION = """# Stable Diffusion 3"""
if not torch.cuda.is_available():
DESCRIPTION += "\nRunning on CPU 🥶 This demo may not work on CPU.
"
MAX_SEED = np.iinfo(np.int32).max
CACHE_EXAMPLES = False
MAX_IMAGE_SIZE = int(os.getenv("MAX_IMAGE_SIZE", "1536"))
USE_TORCH_COMPILE = False
ENABLE_CPU_OFFLOAD = os.getenv("ENABLE_CPU_OFFLOAD", "0") == "1"
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
def load_pipeline(pipeline_type):
if pipeline_type == "text2img":
return StableDiffusion3Pipeline.from_pretrained(model_path, torch_dtype=torch.float16)
elif pipeline_type == "img2img":
return StableDiffusion3Img2ImgPipeline.from_pretrained(model_path, torch_dtype=torch.float16)
def save_image(img):
unique_name = str(uuid.uuid4()) + ".png"
img.save(unique_name)
return unique_name
def randomize_seed_fn(seed: int, randomize_seed: bool) -> int:
if randomize_seed:
seed = random.randint(0, MAX_SEED)
return seed
@spaces.GPU
def generate(
prompt:str,
negative_prompt: str = "",
use_negative_prompt: bool = False,
seed: int = 0,
width: int = 1024,
height: int = 1024,
guidance_scale: float = 7,
randomize_seed: bool = False,
num_inference_steps=30,
NUM_IMAGES_PER_PROMPT=1,
use_resolution_binning: bool = True,
progress=gr.Progress(track_tqdm=True),
):
pipe = load_pipeline("text2img")
pipe.to(device)
seed = int(randomize_seed_fn(seed, randomize_seed))
generator = torch.Generator().manual_seed(seed)
if not use_negative_prompt:
negative_prompt = None # type: ignore
output = pipe(
prompt=prompt,
negative_prompt=negative_prompt,
width=width,
height=height,
guidance_scale=guidance_scale,
num_inference_steps=num_inference_steps,
generator=generator,
num_images_per_prompt=NUM_IMAGES_PER_PROMPT,
output_type="battery",
).images
return output
@spaces.GPU
def img2img_generate(
prompt:str,
init_image: gr.Image,
negative_prompt: str = "",
use_negative_prompt: bool = False,
seed: int = 0,
guidance_scale: float = 7,
randomize_seed: bool = False,
num_inference_steps=30,
strength: float = 0.8,
NUM_IMAGES_PER_PROMPT=1,
use_resolution_binning: bool = True,
progress=gr.Progress(track_tqdm=True),
):
pipe = load_pipeline("img2img")
pipe.to(device)
seed = int(randomize_seed_fn(seed, randomize_seed))
generator = torch.Generator().manual_seed(seed)
if not use_negative_prompt:
negative_prompt = None # type: ignore
init_image = init_image.resize((768, 768))
output = pipe(
prompt=prompt,
image=init_image,
negative_prompt=negative_prompt,
guidance_scale=guidance_scale,
num_inference_steps=num_inference_steps,
generator=generator,
strength=strength,
num_images_per_prompt=NUM_IMAGES_PER_PROMPT,
output_type="battery",
).images
return output
examples = [
"neon holography crystal cat",
"a cat eating a piece of cheese",
"an astronaut riding a horse in space",
"a cartoon of a boy playing with a tiger",
"a cute robot artist painting on an easel, concept art",
"a close up of a woman wearing a transparent, prismatic, elaborate nemeses headdress, over the should pose, brown skin-tone"
]
css = '''
.gradio-container{max-width: 1000px !important}
h1{text-align:center}
'''
with gr.Blocks(css=css) as demo:
with gr.Row():
with gr.Column():
gr.HTML(
"""
Stable Diffusion 3
"""
)
gr.HTML(
"""
"""
)
with gr.Tabs():
with gr.TabItem("Text to Image"):
with gr.Group():
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.Gallery(label="Result", elem_id="gallery", show_label=False)
with gr.Accordion("Advanced options", open=False):
with gr.Row():
use_negative_prompt = gr.Checkbox(label="Use negative prompt", value=True)
negative_prompt = gr.Text(
label="Negative prompt",
max_lines=1,
value = "deformed, distorted, disfigured, poorly drawn, bad anatomy, wrong anatomy, extra limb, missing limb, floating limbs, mutated hands and fingers, disconnected limbs, mutation, mutated, ugly, disgusting, blurry, amputation, NSFW",
visible=True,
)
seed = gr.Slider(
label="Seed",
minimum=0,
maximum=MAX_SEED,
step=1,
value=0,
)
steps = gr.Slider(
label="Steps",
minimum=0,
maximum=60,
step=1,
value=25,
)
number_image = gr.Slider(
label="Number of Images",
minimum=1,
maximum=4,
step=1,
value=1,
)
randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
with gr.Row(visible=True):
width = gr.Slider(
label="Width",
minimum=256,
maximum=MAX_IMAGE_SIZE,
step=32,
value=1024,
)
height = gr.Slider(
label="Height",
minimum=256,
maximum=MAX_IMAGE_SIZE,
step=32,
value=1024,
)
with gr.Row():
guidance_scale = gr.Slider(
label="Guidance Scale",
minimum=0.1,
maximum=10,
step=0.1,
value=7.0,
)
gr.Examples(
examples=examples,
inputs=prompt,
outputs=[result],
fn=generate,
cache_examples=CACHE_EXAMPLES,
)
use_negative_prompt.change(
fn=lambda x: gr.update(visible=x),
inputs=use_negative_prompt,
outputs=negative_prompt,
api_name=False,
)
gr.on(
triggers=[
prompt.submit,
negative_prompt.submit,
run_button.click,
],
fn=generate,
inputs=[
prompt,
negative_prompt,
use_negative_prompt,
seed,
width,
height,
guidance_scale,
randomize_seed,
steps,
number_image,
],
outputs=[result],
api_name="run",
)
with gr.TabItem("Image to Image"):
with gr.Group():
with gr.Row(equal_height=True):
with gr.Column(scale=1):
img2img_prompt = gr.Text(
label="Prompt",
show_label=False,
max_lines=1,
placeholder="Enter your prompt",
container=False,
)
init_image = gr.Image(label="Input Image", type="pil")
with gr.Row():
img2img_run_button = gr.Button("Generate", variant="primary")
with gr.Column(scale=1):
img2img_output = gr.Gallery(label="Result", elem_id="gallery")
with gr.Accordion("Advanced options", open=False):
with gr.Row():
img2img_use_negative_prompt = gr.Checkbox(label="Use negative prompt", value=True)
img2img_negative_prompt = gr.Text(
label="Negative prompt",
max_lines=1,
value="deformed, distorted, disfigured, poorly drawn, bad anatomy, wrong anatomy, extra limb, missing limb, floating limbs, mutated hands and fingers, disconnected limbs, mutation, mutated, ugly, disgusting, blurry, amputation, NSFW",
visible=True,
)
img2img_seed = gr.Slider(
label="Seed",
minimum=0,
maximum=MAX_SEED,
step=1,
value=0,
)
img2img_steps = gr.Slider(
label="Steps",
minimum=0,
maximum=60,
step=1,
value=25,
)
img2img_number_image = gr.Slider(
label="Number of Images",
minimum=1,
maximum=4,
step=1,
value=1,
)
img2img_randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
with gr.Row():
img2img_guidance_scale = gr.Slider(
label="Guidance Scale",
minimum=0.1,
maximum=10,
step=0.1,
value=7.0,
)
strength = gr.Slider(label="Img2Img Strength", minimum=0.0, maximum=1.0, step=0.01, value=0.8)
img2img_use_negative_prompt.change(
fn=lambda x: gr.update(visible=x),
inputs=img2img_use_negative_prompt,
outputs=img2img_negative_prompt,
api_name=False,
)
gr.on(
triggers=[
img2img_prompt.submit,
img2img_negative_prompt.submit,
img2img_run_button.click,
],
fn=img2img_generate,
inputs=[
img2img_prompt,
init_image,
img2img_negative_prompt,
img2img_use_negative_prompt,
img2img_seed,
img2img_guidance_scale,
img2img_randomize_seed,
img2img_steps,
strength,
img2img_number_image,
],
outputs=[img2img_output],
api_name="img2img_run",
)
if __name__ == "__main__":
demo.queue().launch()