اشترك بالحزمة الذهبية واحصل على وصول غير محدود شمرا أكاديميا
تسجيل مستخدم جديدImproved Transformer for High-Resolution GANs
113
0
0.0
(
0
)
اسأل ChatGPT حول البحث
ﻻ يوجد ملخص باللغة العربية
Attention-based models, exemplified by the Transformer, can effectively model long range dependency, but suffer from the quadratic complexity of self-attention operation, making them difficult to be adopted for high-resolution image generation based on Generative Adversarial Networks (GANs). In this paper, we introduce two key ingredients to Transformer to address this challenge. First, in low-resolution stages of the generative process, standard global self-attention is replaced with the proposed multi-axis blocked self-attention which allows efficient mixing of local and global attention. Second, in high-resolution stages, we drop self-attention while only keeping multi-layer perceptrons reminiscent of the implicit neural function. To further improve the performance, we introduce an additional self-modulation component based on cross-attention. The resulting model, denoted as HiT, has a linear computational complexity with respect to the image size and thus directly scales to synthesizing high definition images. We show in the experiments that the proposed HiT achieves state-of-the-art FID scores of 31.87 and 2.95 on unconditional ImageNet $128 times 128$ and FFHQ $256 times 256$, respectively, with a reasonable throughput. We believe the proposed HiT is an important milestone for generators in GANs which are completely free of convolutions.
قيم البحث
اقرأ أيضاً
Image quality measurement is a critical problem for image super-resolution (SR) algorithms. Usually, they are evaluated by some well-known objective metrics, e.g., PSNR and SSIM, but these indices cannot provide suitable results in accordance with th
e perception of human being. Recently, a more reasonable perception measurement has been proposed in [1], which is also adopted by the PIRM-SR 2018 challenge. In this paper, motivated by [1], we aim to generate a high-quality SR result which balances between the two indices, i.e., the perception index and root-mean-square error (RMSE). To do so, we design a new deep SR framework, dubbed Bi-GANs-ST, by integrating two complementary generative adversarial networks (GAN) branches. One is memory residual SRGAN (MR-SRGAN), which emphasizes on improving the objective performance, such as reducing the RMSE. The other is weight perception SRGAN (WP-SRGAN), which obtains the result that favors better subjective perception via a two-stage adversarial training mechanism. Then, to produce final result with excellent perception scores and RMSE, we use soft-thresholding method to merge the results generated by the two GANs. Our method performs well on the perceptual image super-resolution task of the PIRM 2018 challenge. Experimental results on five benchmarks show that our proposal achieves highly competent performance compared with other state-of-the-art methods.
We present a framework to understand GAN training as alternating density ratio estimation and approximate divergence minimization. This provides an interpretation for the mismatched GAN generator and discriminator objectives often used in practice, a
nd explains the problem of poor sample diversity. We also derive a family of generator objectives that target arbitrary $f$-divergences without minimizing a lower bound, and use them to train generative image models that target either improved sample quality or greater sample diversity.
Video super-resolution (VSR), with the aim to restore a high-resolution video from its corresponding low-resolution version, is a spatial-temporal sequence prediction problem. Recently, Transformer has been gaining popularity due to its parallel comp
uting ability for sequence-to-sequence modeling. Thus, it seems to be straightforward to apply the vision Transformer to solve VSR. However, the typical block design of Transformer with a fully connected self-attention layer and a token-wise feed-forward layer does not fit well for VSR due to the following two reasons. First, the fully connected self-attention layer neglects to exploit the data locality because this layer relies on linear layers to compute attention maps. Second, the token-wise feed-forward layer lacks the feature alignment which is important for VSR since this layer independently processes each of the input token embeddings without any interaction among them. In this paper, we make the first attempt to adapt Transformer for VSR. Specifically, to tackle the first issue, we present a spatial-temporal convolutional self-attention layer with a theoretical understanding to exploit the locality information. For the second issue, we design a bidirectional optical flow-based feed-forward layer to discover the correlations across different video frames and also align features. Extensive experiments on several benchmark datasets demonstrate the effectiveness of our proposed method. The code will be available at https://github.com/caojiezhang/VSR-Transformer.
This paper presents a new Vision Transformer (ViT) architecture Multi-Scale Vision Longformer, which significantly enhances the ViT of cite{dosovitskiy2020image} for encoding high-resolution images using two techniques. The first is the multi-scale m
odel structure, which provides image encodings at multiple scales with manageable computational cost. The second is the attention mechanism of vision Longformer, which is a variant of Longformer cite{beltagy2020longformer}, originally developed for natural language processing, and achieves a linear complexity w.r.t. the number of input tokens. A comprehensive empirical study shows that the new ViT significantly outperforms several strong baselines, including the existing ViT models and their ResNet counterparts, and the Pyramid Vision Transformer from a concurrent work cite{wang2021pyramid}, on a range of vision tasks, including image classification, object detection, and segmentation. The models and source code are released at url{https://github.com/microsoft/vision-longformer}.
Single image super-resolution task has witnessed great strides with the development of deep learning. However, most existing studies focus on building a more complex neural network with a massive number of layers, bringing heavy computational cost an
d memory storage. Recently, as Transformer yields brilliant results in NLP tasks, more and more researchers start to explore the application of Transformer in computer vision tasks. But with the heavy computational cost and high GPU memory occupation of the vision Transformer, the network can not be designed too deep. To address this problem, we propose a novel Efficient Super-Resolution Transformer (ESRT) for fast and accurate image super-resolution. ESRT is a hybrid Transformer where a CNN-based SR network is first designed in the front to extract deep features. Specifically, there are two backbones for formatting the ESRT: lightweight CNN backbone (LCB) and lightweight Transformer backbone (LTB). Among them, LCB is a lightweight SR network to extract deep SR features at a low computational cost by dynamically adjusting the size of the feature map. LTB is made up of an efficient Transformer (ET) with a small GPU memory occupation, which benefited from the novel efficient multi-head attention (EMHA). In EMHA, a feature split module (FSM) is proposed to split the long sequence into sub-segments and then these sub-segments are applied by attention operation. This module can significantly decrease the GPU memory occupation. Extensive experiments show that our ESRT achieves competitive results. Compared with the original Transformer which occupies 16057M GPU memory, the proposed ET only occupies 4191M GPU memory with better performance.
سجل دخول لتتمكن من نشر تعليقات
التعليقات
جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
