Salvato in:
Dettagli Bibliografici
Autori principali: Palit, Sanchar, Dendi, Sathya Veera Reddy, Talluri, Mallikarjuna, Gadde, Raj Narayana
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:https://arxiv.org/abs/2411.06119
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866912569784008704
author Palit, Sanchar
Dendi, Sathya Veera Reddy
Talluri, Mallikarjuna
Gadde, Raj Narayana
author_facet Palit, Sanchar
Dendi, Sathya Veera Reddy
Talluri, Mallikarjuna
Gadde, Raj Narayana
contents Vision Transformers and U-Net architectures have been widely adopted in the implementation of Diffusion Models. However, each architecture presents specific challenges while realizing them on-device. Vision Transformers require positional embedding to maintain correspondence between the tokens processed by the transformer, although they offer the advantage of using fixed-size, reusable repetitive blocks following tokenization. The U-Net architecture lacks these attributes, as it utilizes variable-sized intermediate blocks for down-convolution and up-convolution in the noise estimation backbone for the diffusion process. To address these issues, we propose an architecture that utilizes a fixed-size, reusable transformer block as a core structure, making it more suitable for hardware implementation. Our architecture is characterized by low complexity, token-free design, absence of positional embeddings, uniformity, and scalability, making it highly suitable for deployment on mobile and resource-constrained devices. The proposed model exhibit competitive and consistent performance across both unconditional and conditional image generation tasks. The model achieved a state-of-the-art FID score of 1.6 on unconditional image generation with the CelebA.
format Preprint
id arxiv_https___arxiv_org_abs_2411_06119
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Hardware-Friendly Diffusion Models with Fixed-Size Reusable Structures for On-Device Image Generation
Palit, Sanchar
Dendi, Sathya Veera Reddy
Talluri, Mallikarjuna
Gadde, Raj Narayana
Computer Vision and Pattern Recognition
Machine Learning
Vision Transformers and U-Net architectures have been widely adopted in the implementation of Diffusion Models. However, each architecture presents specific challenges while realizing them on-device. Vision Transformers require positional embedding to maintain correspondence between the tokens processed by the transformer, although they offer the advantage of using fixed-size, reusable repetitive blocks following tokenization. The U-Net architecture lacks these attributes, as it utilizes variable-sized intermediate blocks for down-convolution and up-convolution in the noise estimation backbone for the diffusion process. To address these issues, we propose an architecture that utilizes a fixed-size, reusable transformer block as a core structure, making it more suitable for hardware implementation. Our architecture is characterized by low complexity, token-free design, absence of positional embeddings, uniformity, and scalability, making it highly suitable for deployment on mobile and resource-constrained devices. The proposed model exhibit competitive and consistent performance across both unconditional and conditional image generation tasks. The model achieved a state-of-the-art FID score of 1.6 on unconditional image generation with the CelebA.
title Hardware-Friendly Diffusion Models with Fixed-Size Reusable Structures for On-Device Image Generation
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2411.06119