【Arc A770】Z Image Turboを使ってみる（NVlabs/Sana と比べてみました）

はじめに

Z Image Truboに関してはこちら。 github.com

NVlabs/Sanaに関してはこちら。 github.com

結果

Z Image Turboの結果

NVlabs/Sanaの結果

「HuggingFace」の文字が破綻しています。

しかも文字が浮いています。

Pythonスクリプト

Z Image Turbo

import os

os.environ["UR_L0_ENABLE_RELAXED_ALLOCATION_LIMITS"] = "1"

import torch
from diffusers import ZImagePipeline

pipe = ZImagePipeline.from_pretrained(
    "Tongyi-MAI/Z-Image-Turbo",
    torch_dtype=torch.bfloat16
)

pipe.enable_model_cpu_offload(device="xpu")

prompt = 'Ultra-realistic, high-quality photo of an anthropomorphic capybara with a tough, streetwise attitude, wearing a worn black leather jacket, dark sunglasses, and ripped jeans. The capybara is leaning casually against a gritty urban wall covered in vibrant graffiti. Behind it, in bold, dripping yellow spray paint, the word "HuggingFace" is scrawled in large street-art style letters. The scene is set in a dimly lit alleyway with moody lighting, scattered trash, and an edgy, rebellious vibe — like a character straight out of an underground comic book.'

image = pipe(
    prompt,
    height=1024,
    width=1024,
    num_inference_steps=9,
    guidance_scale=0.0,
    generator=torch.manual_seed(42),
).images[0]

image.save("zimage.png")

Sana

import torch
from diffusers import SanaPAGPipeline

pipe = SanaPAGPipeline.from_pretrained(
    "Efficient-Large-Model/Sana_1600M_1024px_BF16_diffusers",
    variant="bf16",
    torch_dtype=torch.bfloat16,
    pag_applied_layers="transformer_blocks.8",
).to("xpu")

prompt = 'Ultra-realistic, high-quality photo of an anthropomorphic capybara with a tough, streetwise attitude, wearing a worn black leather jacket, dark sunglasses, and ripped jeans. The capybara is leaning casually against a gritty urban wall covered in vibrant graffiti. Behind it, in bold, dripping yellow spray paint, the word "HuggingFace" is scrawled in large street-art style letters. The scene is set in a dimly lit alleyway with moody lighting, scattered trash, and an edgy, rebellious vibe — like a character straight out of an underground comic book.'

image = pipe(
    prompt=prompt,
    guidance_scale=5.0,
    pag_scale=2.0,
    num_inference_steps=20,
    generator=torch.manual_seed(42),
)[0]

image[0].save("with_pag.jpg")

環境構築

それぞれのpyproject.tomlを示します。

Z Image Turbo

[project]
name = "zimage"
version = "0.1.0"
description = "Add your description here"
readme = "README.md"
requires-python = ">=3.13"
dependencies = [
    "accelerate==1.12.0",
    "diffusers==0.36.0",
    "hf-xet==1.2.0",
    "pytorch-triton-xpu==3.5.0",
    "torch==2.9.1+xpu",
    "transformers==4.57.3",
]

[[tool.uv.index]]
name = "torch-xpu"
url = "https://download.pytorch.org/whl/xpu"
explicit = true

[tool.uv.sources]
torch = [{ index = "torch-xpu" }]
pytorch-triton-xpu = [{ index = "torch-xpu" }]

Sana

[project]
name = "sana"
version = "0.1.0"
description = "Add your description here"
readme = "README.md"
requires-python = ">=3.13"
dependencies = [
    "accelerate==1.12.0",
    "beautifulsoup4==4.14.3",
    "diffusers==0.35.2",
    "ftfy==6.3.1",
    "hf-xet==1.2.0",
    "pytorch-triton-xpu==3.5.0",
    "torch==2.9.1+xpu",
    "transformers==4.57.3",
]

[[tool.uv.index]]
name = "torch-xpu"
url = "https://download.pytorch.org/whl/xpu"
explicit = true

[tool.uv.sources]
torch = [{ index = "torch-xpu" }]
pytorch-triton-xpu = [{ index = "torch-xpu" }]