Dc Ae F32c32 Sana 1.1

📚 摘要

我们提出了深度压缩自编码器（Deep Compression Autoencoder，DC - AE），这是一个用于加速高分辨率扩散模型的新型自编码器模型家族。现有的自编码器模型在中等空间压缩比（如 8 倍）下取得了令人印象深刻的结果，但在高空间压缩比（如 64 倍）下无法保持令人满意的重建精度。我们通过引入两项关键技术来应对这一挑战：（1）残差自编码，我们设计模型基于空间到通道转换后的特征学习残差，以缓解高空间压缩自编码器的优化难度；（2）解耦高分辨率自适应，一种高效的解耦三阶段训练策略，用于减轻高空间压缩自编码器的泛化惩罚。通过这些设计，我们将自编码器的空间压缩比提高到 128，同时保持了重建质量。将我们的 DC - AE 应用于潜在扩散模型，我们在不降低精度的情况下实现了显著的加速。例如，在 ImageNet 512x512 上，与广泛使用的 SD - VAE - f8 自编码器相比，我们的 DC - AE 在 H100 GPU 上为 UViT - H 提供了 19.1 倍的推理加速和 17.9 倍的训练加速，同时实现了更好的 FID。

💻 使用示例

基础用法

# build DC-AE models
# full DC-AE model list: https://huggingface.co/collections/mit-han-lab/dc-ae-670085b9400ad7197bb1009b
from efficientvit.ae_model_zoo import DCAE_HF

dc_ae = DCAE_HF.from_pretrained(f"mit-han-lab/dc-ae-f64c128-in-1.0")

# encode
from PIL import Image
import torch
import torchvision.transforms as transforms
from torchvision.utils import save_image
from efficientvit.apps.utils.image import DMCrop

device = torch.device("cuda")
dc_ae = dc_ae.to(device).eval()

transform = transforms.Compose([
    DMCrop(512), # resolution
    transforms.ToTensor(),
    transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)),
])
image = Image.open("assets/fig/girl.png")
x = transform(image)[None].to(device)
latent = dc_ae.encode(x)
print(latent.shape)

# decode
y = dc_ae.decode(latent)
save_image(y * 0.5 + 0.5, "demo_dc_ae.png")

高级用法

# build DC-AE-Diffusion models
# full DC-AE-Diffusion model list: https://huggingface.co/collections/mit-han-lab/dc-ae-diffusion-670dbb8d6b6914cf24c1a49d
from efficientvit.diffusion_model_zoo import DCAE_Diffusion_HF

dc_ae_diffusion = DCAE_Diffusion_HF.from_pretrained(f"mit-han-lab/dc-ae-f64c128-in-1.0-uvit-h-in-512px-train2000k")

# denoising on the latent space
import torch
import numpy as np
from torchvision.utils import save_image

torch.set_grad_enabled(False)
device = torch.device("cuda")
dc_ae_diffusion = dc_ae_diffusion.to(device).eval()

seed = 0
torch.manual_seed(seed)
torch.cuda.manual_seed_all(seed)
eval_generator = torch.Generator(device=device)
eval_generator.manual_seed(seed)

prompts = torch.tensor(
    [279, 333, 979, 936, 933, 145, 497, 1, 248, 360, 793, 12, 387, 437, 938, 978], dtype=torch.int, device=device
)
num_samples = prompts.shape[0]
prompts_null = 1000 * torch.ones((num_samples,), dtype=torch.int, device=device)
latent_samples = dc_ae_diffusion.diffusion_model.generate(prompts, prompts_null, 6.0, eval_generator)
latent_samples = latent_samples / dc_ae_diffusion.scaling_factor

# decode
image_samples = dc_ae_diffusion.autoencoder.decode(latent_samples)
save_image(image_samples * 0.5 + 0.5, "demo_dc_ae_diffusion.png", nrow=int(np.sqrt(num_samples)))

📖 引用

如果 DC - AE 对你的研究有用或相关，请通过引用我们的论文来认可我们的贡献：

@article{chen2024deep,
  title={Deep Compression Autoencoder for Efficient High-Resolution Diffusion Models},
  author={Chen, Junyu and Cai, Han and Chen, Junsong and Xie, Enze and Yang, Shang and Tang, Haotian and Li, Muyang and Lu, Yao and Han, Song},
  journal={arXiv preprint arXiv:2410.10733},
  year={2024}
}

关于 SANA 1.5 的工作可以按如下方式引用：

@misc{xie2025sana,
      title={SANA 1.5: Efficient Scaling of Training-Time and Inference-Time Compute in Linear Diffusion Transformer},
      author={Enze Xie and Junyu Chen and Han Cai and Junsong Chen and Haotian Tang and Yao Lu and Song Han},
      year={2025},
      eprint={2501.18427},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2501.18427},
}