Spaces:
Running
on
Zero
Running
on
Zero
File size: 3,851 Bytes
45bea6f f892d4a 45bea6f f892d4a 45bea6f d1eabfd 45bea6f d1eabfd d0b932a 45bea6f d1eabfd d0b932a 45bea6f d0b932a d1eabfd d0b932a 45bea6f d1eabfd 45bea6f b7bf69e 45bea6f b7bf69e 45bea6f d0b932a 45bea6f d1eabfd 45bea6f a3b0b89 45bea6f d1eabfd 45bea6f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 |
import gradio as gr
import numpy as np
import random
from diffusers import PixArtAlphaPipeline, Transformer2DModel, LCMScheduler
import torch
from peft import PeftModel
device = "cuda" if torch.cuda.is_available() else "cpu"
transformer = Transformer2DModel.from_pretrained(
"PixArt-alpha/PixArt-XL-2-1024-MS",
subfolder="transformer",
torch_dtype=torch.float16
)
transformer = PeftModel.from_pretrained(
transformer,
"jasperai/flash-pixart"
)
if torch.cuda.is_available():
torch.cuda.max_memory_allocated(device=device)
pipe = PixArtAlphaPipeline.from_pretrained(
"PixArt-alpha/PixArt-XL-2-1024-MS",
transformer=transformer,
torch_dtype=torch.float16
)
pipe.enable_xformers_memory_efficient_attention()
pipe = pipe.to(device)
else:
pipe = PixArtAlphaPipeline.from_pretrained(
"PixArt-alpha/PixArt-XL-2-1024-MS",
transformer=transformer,
torch_dtype=torch.float16
)
pipe = pipe.to(device)
pipe.scheduler = LCMScheduler.from_pretrained(
"PixArt-alpha/PixArt-XL-2-1024-MS",
subfolder="scheduler",
timestep_spacing="trailing",
)
MAX_SEED = np.iinfo(np.int32).max
MAX_IMAGE_SIZE = 1024
NUM_INFERENCE_STEPS = 4
def infer(prompt, seed, randomize_seed, num_inference_steps):
if randomize_seed:
seed = random.randint(0, MAX_SEED)
generator = torch.Generator().manual_seed(seed)
image = pipe(
prompt = prompt,
guidance_scale = 0,
num_inference_steps = num_inference_steps,
generator = generator
).images[0]
return image
examples = [
"The image showcases a freshly baked bread, possibly focaccia, with rosemary sprigs and red pepper flakes sprinkled on top. It's sliced and placed on a wire cooling rack, with a bowl of mixed peppercorns beside it.",
"A raccoon reading a book in a lush forest.",
"A serene landscape showcases a winding road alongside a vast, turquoise lake, flanked by majestic snow-capped mountains under a partly cloudy sky.",
]
css="""
#col-container {
margin: 0 auto;
max-width: 512px;
}
"""
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"""
# ⚡ FlashDiffusion: FlashPixart ⚡
This is an interactive demo of [Flash Diffusion](https://huggingface.co/jasperai/flash-pixart), a diffusion distillation method proposed in [ADD ARXIV]() *by Clément Chadebec, Onur Tasar and Benjamin Aubin.*
This model is a **66.5M** LoRA distilled version of [Pixart-α](https://huggingface.co/PixArt-alpha/PixArt-XL-2-1024-MS) model that is able to generate 1024x1024 images in **4 steps**.
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):
seed = gr.Slider(
label="Seed",
minimum=0,
maximum=MAX_SEED,
step=1,
value=0,
)
randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
gr.Examples(
examples = examples,
inputs = [prompt]
)
run_button.click(
fn = infer,
inputs = [prompt, seed, randomize_seed, NUM_INFERENCE_STEPS],
outputs = [result]
)
demo.queue().launch() |