Running on CPU 🥶 This demo does not work on CPU.
"
IS_COLAB = utils.is_google_colab() or os.getenv("IS_COLAB") == "1"
HF_TOKEN = os.getenv("HF_TOKEN")
CACHE_EXAMPLES = torch.cuda.is_available() and os.getenv("CACHE_EXAMPLES") == "1"
MIN_IMAGE_SIZE = int(os.getenv("MIN_IMAGE_SIZE", "512"))
MAX_IMAGE_SIZE = int(os.getenv("MAX_IMAGE_SIZE", "2048"))
USE_TORCH_COMPILE = os.getenv("USE_TORCH_COMPILE") == "1"
ENABLE_CPU_OFFLOAD = os.getenv("ENABLE_CPU_OFFLOAD") == "1"
OUTPUT_DIR = os.getenv("OUTPUT_DIR", "./outputs")
HISTORY_SECRET = os.getenv("HISTORY_SECRET", "default_secret")
MODEL = os.getenv(
"MODEL",
"https://huggingface.co/AstraliteHeart/pony-diffusion-v6/blob/main/v6.safetensors",
)
DESCRIPTION = '''
High Definition Pony Diffusion
Gradio demo for PonyDiffusion v6 with image gallery, json prompt support, advanced options and more.
❤️ Thanks for ✨10k visits! FLUX INTEGRATION COMING SOON!
🔎 For more details about me, take a look at My website.
🌚 For dark mode compatibility, click here.
'''
torch.backends.cudnn.deterministic = True
torch.backends.cudnn.benchmark = False
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
def load_pipeline(model_name):
vae = AutoencoderKL.from_pretrained(
"madebyollin/sdxl-vae-fp16-fix",
torch_dtype=torch.float16,
)
pipeline = (
StableDiffusionXLPipeline.from_single_file
if MODEL.endswith(".safetensors")
else StableDiffusionXLPipeline.from_pretrained
)
pipe = pipeline(
model_name,
vae=vae,
torch_dtype=torch.float16,
custom_pipeline="lpw_stable_diffusion_xl",
use_safetensors=True,
add_watermarker=False,
use_auth_token=HF_TOKEN,
variant="fp16",
)
pipe.to(device)
return pipe
def parse_json_parameters(json_str):
try:
params = json.loads(json_str)
required_keys = ['prompt', 'negative_prompt', 'resolution', 'guidance_scale', 'num_inference_steps', 'seed', 'sampler']
for key in required_keys:
if key not in params:
raise ValueError(f"Missing required key: {key}")
width, height = map(int, params['resolution'].split(' x '))
return {
'prompt': params['prompt'],
'negative_prompt': params['negative_prompt'],
'seed': params['seed'],
'width': width,
'height': height,
'guidance_scale': params['guidance_scale'],
'num_inference_steps': params['num_inference_steps'],
'sampler': params['sampler'],
'use_upscaler': params.get('use_upscaler', False)
}
except json.JSONDecodeError:
raise ValueError("Invalid JSON format")
except Exception as e:
raise ValueError(f"Error parsing JSON: {str(e)}")
@spaces.GPU
def generate(
prompt: str,
negative_prompt: str = "",
seed: int = 0,
custom_width: int = 1024,
custom_height: int = 1024,
guidance_scale: float = 7.0,
num_inference_steps: int = 30,
sampler: str = "DPM++ 2M SDE Karras",
aspect_ratio_selector: str = "1024 x 1024",
use_upscaler: bool = False,
upscaler_strength: float = 0.55,
upscale_by: float = 1.5,
json_params: str = "",
batch_size: int = 1,
progress=gr.Progress(track_tqdm=True),
) -> Image:
if json_params:
try:
params = parse_json_parameters(json_params)
prompt = params['prompt']
negative_prompt = params['negative_prompt']
seed = params['seed']
custom_width = params['width']
custom_height = params['height']
guidance_scale = params['guidance_scale']
num_inference_steps = params['num_inference_steps']
sampler = params['sampler']
use_upscaler = params['use_upscaler']
except ValueError as e:
raise gr.Error(str(e))
generator = utils.seed_everything(seed)
width, height = utils.aspect_ratio_handler(
aspect_ratio_selector,
custom_width,
custom_height,
)
width, height = utils.preprocess_image_dimensions(width, height)
backup_scheduler = pipe.scheduler
pipe.scheduler = utils.get_scheduler(pipe.scheduler.config, sampler)
if use_upscaler:
upscaler_pipe = StableDiffusionXLImg2ImgPipeline(**pipe.components)
metadata = {
"prompt": prompt,
"negative_prompt": negative_prompt,
"resolution": f"{width} x {height}",
"guidance_scale": guidance_scale,
"num_inference_steps": num_inference_steps,
"seed": seed,
"sampler": sampler,
"batch_size": batch_size,
}
if use_upscaler:
new_width = int(width * upscale_by)
new_height = int(height * upscale_by)
metadata["use_upscaler"] = {
"upscale_method": "nearest-exact",
"upscaler_strength": upscaler_strength,
"upscale_by": upscale_by,
"new_resolution": f"{new_width} x {new_height}",
}
else:
metadata["use_upscaler"] = None
logger.info(json.dumps(metadata, indent=4))
try:
all_images = []
for _ in range(batch_size):
batch_generator = utils.seed_everything(random.randint(0, utils.MAX_SEED))
if use_upscaler:
latents = pipe(
prompt=prompt,
negative_prompt=negative_prompt,
width=width,
height=height,
guidance_scale=guidance_scale,
num_inference_steps=num_inference_steps,
generator=batch_generator,
output_type="latent",
).images
upscaled_latents = utils.upscale(latents, "nearest-exact", upscale_by)
images = upscaler_pipe(
prompt=prompt,
negative_prompt=negative_prompt,
image=upscaled_latents,
guidance_scale=guidance_scale,
num_inference_steps=num_inference_steps,
strength=upscaler_strength,
generator=batch_generator,
output_type="pil",
).images
else:
images = pipe(
prompt=prompt,
negative_prompt=negative_prompt,
width=width,
height=height,
guidance_scale=guidance_scale,
num_inference_steps=num_inference_steps,
generator=batch_generator,
output_type="pil",
).images
all_images.extend(images)
if all_images and IS_COLAB:
for image in all_images:
filepath = utils.save_image(image, metadata, OUTPUT_DIR)
logger.info(f"Image saved as {filepath} with metadata")
return all_images, metadata
except Exception as e:
logger.exception(f"An error occurred: {e}")
raise
finally:
if use_upscaler:
del upscaler_pipe
pipe.scheduler = backup_scheduler
utils.free_memory()
generation_history = []
def update_history_list():
return [item["image"] for item in generation_history]
def handle_image_click(evt: gr.SelectData):
selected = generation_history[evt.index]
return selected["image"], json.dumps(selected["metadata"], indent=2)
def generate_and_update_history(*args, **kwargs):
global generation_history
images, metadata = generate(*args, **kwargs)
timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
for image in images:
generation_history.insert(0, {
"prompt": metadata["prompt"],
"timestamp": timestamp,
"image": image,
"metadata": metadata
})
if len(generation_history) > 20:
generation_history = generation_history[:20]
return images[0], json.dumps(metadata, indent=2), update_history_list()
with open('characterfull.txt', 'r') as f:
characters = [line.strip() for line in f.readlines()]
def get_random_character():
return random.choice(characters)
def add_quality_tags(prompt, negative_prompt):
positive_tags = "score_9, score_8_up, score_7_up, score_6_up, dramatic lighting"
negative_tags = "worst quality, low quality, text, censored, deformed, bad hand, blurry, (watermark), mutated hands, monochrome"
new_prompt = f"{positive_tags}, {prompt}" if prompt else positive_tags
new_negative_prompt = f"{negative_tags}, {negative_prompt}" if negative_prompt else negative_tags
return new_prompt, new_negative_prompt
if torch.cuda.is_available():
pipe = load_pipeline(MODEL)
logger.info("Loaded on Device!")
else:
pipe = None
def check_history_password(password):
if password == HISTORY_SECRET:
return gr.update(visible=True)
else:
return gr.update(visible=False)
with gr.Blocks(css="style.css") as demo:
gr.Markdown(DESCRIPTION)
gr.DuplicateButton(
value="Duplicate Space for private use",
elem_id="duplicate-button",
visible=os.getenv("SHOW_DUPLICATE_BUTTON") == "1",
)
with gr.Group():
with gr.Row():
prompt = gr.Text(
label="Prompt",
show_label=False,
max_lines=5,
placeholder="Enter your prompt",
container=False,
)
run_button = gr.Button(
"Generate",
variant="primary",
scale=0
)
with gr.Row():
random_character_button = gr.Button("Random Character")
add_quality_tags_button = gr.Button("Add quality tags")
result = gr.Image(
label="Result",
show_label=False
)
with gr.Accordion(label="Advanced Settings", open=False):
negative_prompt = gr.Text(
label="Negative Prompt",
max_lines=5,
placeholder="Enter a negative prompt",
value=""
)
aspect_ratio_selector = gr.Radio(
label="Aspect Ratio",
choices=config.aspect_ratios,
value="1024 x 1024",
container=True,
)
with gr.Group(visible=False) as custom_resolution:
with gr.Row():
custom_width = gr.Slider(
label="Width",
minimum=MIN_IMAGE_SIZE,
maximum=MAX_IMAGE_SIZE,
step=8,
value=1024,
)
custom_height = gr.Slider(
label="Height",
minimum=MIN_IMAGE_SIZE,
maximum=MAX_IMAGE_SIZE,
step=8,
value=1024,
)
use_upscaler = gr.Checkbox(label="Use Upscaler", value=False)
with gr.Row() as upscaler_row:
upscaler_strength = gr.Slider(
label="Strength",
minimum=0,
maximum=1,
step=0.05,
value=0.55,
visible=False,
)
upscale_by = gr.Slider(
label="Upscale by",
minimum=1,
maximum=1.5,
step=0.1,
value=1.5,
visible=False,
)
sampler = gr.Dropdown(
label="Sampler",
choices=config.sampler_list,
interactive=True,
value="DPM++ 2M SDE Karras",
)
with gr.Row():
seed = gr.Slider(
label="Seed", minimum=0, maximum=utils.MAX_SEED, step=1, value=0
)
randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
with gr.Group():
with gr.Row():
guidance_scale = gr.Slider(
label="Guidance scale",
minimum=1,
maximum=12,
step=0.1,
value=7.0,
)
num_inference_steps = gr.Slider(
label="Number of inference steps",
minimum=1,
maximum=50,
step=1,
value=28,
)
batch_size = gr.Slider(
label="Batch Size",
minimum=1,
maximum=4,
step=1,
value=1,
)
with gr.Accordion(label="Generation Parameters", open=False):
gr_metadata = gr.JSON(label="Metadata", show_label=False)
json_input = gr.TextArea(label="Edit/Paste JSON Parameters", placeholder="Paste or edit JSON parameters here")
generate_from_json = gr.Button("Generate from JSON")
with gr.Accordion("Generation History", open=False) as history_accordion:
history_password = gr.Textbox(
label="Enable generation history; do not generate illegal or harmful content.",
type="password",
placeholder="GLOBAL GENERATION HISTORY IS DISABLED"
)
history_submit = gr.Button("Submit")
with gr.Group(visible=False) as history_content:
history_gallery = gr.Gallery(
label="History",
show_label=False,
elem_id="history_gallery",
columns=5,
rows=2,
height="auto"
)
with gr.Row():
selected_image = gr.Image(label="Selected Image", interactive=False)
selected_metadata = gr.JSON(label="Selected Metadata", show_label=False)
gr.Examples(
examples=config.examples,
inputs=prompt,
outputs=[result, gr_metadata],
fn=lambda *args, **kwargs: generate_and_update_history(*args, use_upscaler=True, **kwargs),
cache_examples=CACHE_EXAMPLES,
)
use_upscaler.change(
fn=lambda x: [gr.update(visible=x), gr.update(visible=x)],
inputs=use_upscaler,
outputs=[upscaler_strength, upscale_by],
queue=False,
api_name=False,
)
aspect_ratio_selector.change(
fn=lambda x: gr.update(visible=x == "Custom"),
inputs=aspect_ratio_selector,
outputs=custom_resolution,
queue=False,
api_name=False,
)
inputs = [
prompt,
negative_prompt,
seed,
custom_width,
custom_height,
guidance_scale,
num_inference_steps,
sampler,
aspect_ratio_selector,
use_upscaler,
upscaler_strength,
upscale_by,
json_input,
batch_size,
]
prompt.submit(
fn=utils.randomize_seed_fn,
inputs=[seed, randomize_seed],
outputs=seed,
queue=False,
api_name=False,
).then(
fn=generate_and_update_history,
inputs=inputs,
outputs=[result, gr_metadata, history_gallery],
)
negative_prompt.submit(
fn=utils.randomize_seed_fn,
inputs=[seed, randomize_seed],
outputs=seed,
queue=False,
api_name=False,
).then(
fn=generate_and_update_history,
inputs=inputs,
outputs=[result, gr_metadata, history_gallery],
)
run_button.click(
fn=utils.randomize_seed_fn,
inputs=[seed, randomize_seed],
outputs=seed,
queue=False,
api_name=False,
).then(
fn=generate_and_update_history,
inputs=inputs,
outputs=[result, gr_metadata, history_gallery],
)
generate_from_json.click(
fn=generate_and_update_history,
inputs=inputs,
outputs=[result, gr_metadata, history_gallery],
)
random_character_button.click(
fn=get_random_character,
inputs=[],
outputs=[prompt]
)
add_quality_tags_button.click(
fn=add_quality_tags,
inputs=[prompt, negative_prompt],
outputs=[prompt, negative_prompt]
)
history_gallery.select(
fn=handle_image_click,
inputs=[],
outputs=[selected_image, selected_metadata]
)
history_submit.click(
fn=check_history_password,
inputs=[history_password],
outputs=[history_content],
)
demo.queue(max_size=20).launch(debug=IS_COLAB, share=IS_COLAB)