Peter Vajda
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Peter Vajda has a total of 30 co-authors. Here are the top co-authors by number of shared publications:
- Bichen Wu
Meta, San Francisco, United States
2 shared publications Lutz Goldmann
Ecole Polytechnique Fédérale de Lausanne, Lausanne, Switzerland
2 shared publications- Peizhao Zhang
Meta, San Francisco United States
2 shared publications - Touradj Ebrahimi
École Polytechnique Fédérale de Lausanne
2 shared publications - Xiaoliang Dai
Meta
2 shared publications - Alvin Wan
UC Berkeley
1 shared publication
Latest Publications
- Open Access
- 26 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 30-38
Text-to-Image Generation Post-Training with Pixel-Space Loss
- 26 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 30-38
Open AccessText-to-Image Generation Post-Training with Pixel-Space Loss
- Open Access
- 11 May 2024
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Embedded Computing Systems
- Vol. 23 (3) , 1-19
An Investigation on Hardware-Aware Vision Transformer Scaling
Vision Transformer (ViT) has demonstrated promising performance in various computer vision tasks, and recently attracted a lot of research attention. Many recent works have focused on proposing new architectures to improve ViT and deploying it into real-world applications. However, little effort has been made to analyze and understand ViT’s architecture design space and its implication for hardware costs on different devices. In this work, by simply scaling ViT’s depth, width, input size, and other basic configurations, we show that a scaled vanilla ViT model without bells and whistles can achieve comparable or superior accuracy-efficiency trade-off than most of the latest ViT variants. Specifically, compared with DeiT-Tiny, our scaled model achieves a ↑ 1.9% higher ImageNet top-1 accuracy under the same FLOPs and a ↑ 3.7% better ImageNet top-1 accuracy under the same latency on an NVIDIA Edge GPU TX2. Motivated by this, we further investigate the extracted scaling strategies from the following two aspects: (1) can these scaling strategies be transferred across different real hardware devices ? and (2) can these scaling strategies be transferred to different ViT variants and tasks ?. For (1), our exploration, based on various devices with different resource budgets, indicates that the transferability effectiveness depends on the underlying device together with its corresponding deployment tool. For (2), we validate the effective transferability of the aforementioned scaling strategies obtained from a vanilla ViT model on top of an image classification task to the PiT model, a strong ViT variant targeting efficiency as well as object detection and video classification tasks. In particular, when transferred to PiT, our scaling strategies lead to a boosted ImageNet top-1 accuracy of from 74.6% to 76.7% (↑ 2.1%) under the same 0.7G FLOPs. When transferred to the COCO object detection task, the average precision is boosted by ↑ 0.7% under a similar throughput on a V100 GPU.
- 11 May 2024
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Embedded Computing Systems
- Vol. 23 (3) , 1-19
Open AccessAn Investigation on Hardware-Aware Vision Transformer Scaling
Vision Transformer (ViT) has demonstrated promising performance in various computer vision tasks, and recently attracted a lot of research attention. Many recent works have focused on proposing new architectures to improve ViT and deploying it into real-world applications. However, little effort has been made to analyze and understand ViT’s architecture design space and its implication for hardware costs on different devices. In this work, by simply scaling ViT’s depth, width, input size, and other basic configurations, we show that a scaled vanilla ViT model without bells and whistles can achieve comparable or superior accuracy-efficiency trade-off than most of the latest ViT variants. Specifically, compared with DeiT-Tiny, our scaled model achieves a ↑ 1.9% higher ImageNet top-1 accuracy under the same FLOPs and a ↑ 3.7% better ImageNet top-1 accuracy under the same latency on an NVIDIA Edge GPU TX2. Motivated by this, we further investigate the extracted scaling strategies from the following two aspects: (1) can these scaling strategies be transferred across different real hardware devices ? and (2) can these scaling strategies be transferred to different ViT variants and tasks ?. For (1), our exploration, based on various devices with different resource budgets, indicates that the transferability effectiveness depends on the underlying device together with its corresponding deployment tool. For (2), we validate the effective transferability of the aforementioned scaling strategies obtained from a vanilla ViT model on top of an image classification task to the PiT model, a strong ViT variant targeting efficiency as well as object detection and video classification tasks. In particular, when transferred to PiT, our scaling strategies lead to a boosted ImageNet top-1 accuracy of from 74.6% to 76.7% (↑ 2.1%) under the same 0.7G FLOPs. When transferred to the COCO object detection task, the average precision is boosted by ↑ 0.7% under a similar throughput on a V100 GPU.
- Open Access
FBNetV3: Joint Architecture-Recipe Search using Predictor Pretraining
Neural Architecture Search (NAS) yields state-of-the-art neural networks that outperform their best manually-designed counterparts. However, previous NAS methods search for architectures under one set of training hyper-parameters (i.e., a training recipe), overlooking superior architecture-recipe combinations. To address this, we present Neural Architecture-Recipe Search (NARS) to search both (a) architectures and (b) their corresponding training recipes, simultaneously. NARS utilizes an accuracy predictor that scores architecture and training recipes jointly, guiding both sample selection and ranking. Furthermore, to compensate for the enlarged search space, we leverage "free" architecture statistics (e.g., FLOP count) to pretrain the predictor, significantly improving its sample efficiency and prediction reliability. After training the predictor via constrained iterative optimization, we run fast evolutionary searches in just CPU minutes to generate architecture-recipe pairs for a variety of resource constraints, called FBNetV3. FBNetV3 makes up a family of state-of-the-art compact neural networks that outperform both automatically and manually-designed competitors. For example, FBNetV3 matches both EfficientNet and ResNeSt accuracy on ImageNet with up to 2.0x and 7.1x fewer FLOPs, respectively. Furthermore, FBNetV3 yields significant performance gains for downstream object detection tasks, improving mAP despite 18% fewer FLOPs and 34% fewer parameters than EfficientNet-based equivalents.
Open AccessFBNetV3: Joint Architecture-Recipe Search using Predictor Pretraining
Neural Architecture Search (NAS) yields state-of-the-art neural networks that outperform their best manually-designed counterparts. However, previous NAS methods search for architectures under one set of training hyper-parameters (i.e., a training recipe), overlooking superior architecture-recipe combinations. To address this, we present Neural Architecture-Recipe Search (NARS) to search both (a) architectures and (b) their corresponding training recipes, simultaneously. NARS utilizes an accuracy predictor that scores architecture and training recipes jointly, guiding both sample selection and ranking. Furthermore, to compensate for the enlarged search space, we leverage "free" architecture statistics (e.g., FLOP count) to pretrain the predictor, significantly improving its sample efficiency and prediction reliability. After training the predictor via constrained iterative optimization, we run fast evolutionary searches in just CPU minutes to generate architecture-recipe pairs for a variety of resource constraints, called FBNetV3. FBNetV3 makes up a family of state-of-the-art compact neural networks that outperform both automatically and manually-designed competitors. For example, FBNetV3 matches both EfficientNet and ResNeSt accuracy on ImageNet with up to 2.0x and 7.1x fewer FLOPs, respectively. Furthermore, FBNetV3 yields significant performance gains for downstream object detection tasks, improving mAP despite 18% fewer FLOPs and 34% fewer parameters than EfficientNet-based equivalents.
- Open Access
- 29 March 2010
- Published by Association for Computing Machinery (ACM)
- p. 497-506
Object-based tag propagation for semi-automatic annotation of images
- 29 March 2010
- Published by Association for Computing Machinery (ACM)
- p. 497-506
Open AccessObject-based tag propagation for semi-automatic annotation of images
- Open Access
- 1 January 2009
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 600-605
Analysis of the Limits of Graph-Based Object Duplicate Detection
Several applications require accurate and efficient object duplicate detection methods, such as automatic video and image tag propagation, video surveillance, and high level image or video search. In this paper, we explore the limits of our recently proposed graph-based object duplicate detection method. The dependency of the performance with respect to the number of training images is assessed and the optimal detection parameters are determined. Furthermore, the differences among various object classes are analyzed. In this way, this paper provides an in-depth analysis of the graph based object duplicate detection method.
- 1 January 2009
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 600-605
Open AccessAnalysis of the Limits of Graph-Based Object Duplicate Detection
Several applications require accurate and efficient object duplicate detection methods, such as automatic video and image tag propagation, video surveillance, and high level image or video search. In this paper, we explore the limits of our recently proposed graph-based object duplicate detection method. The dependency of the performance with respect to the number of training images is assessed and the optimal detection parameters are determined. Furthermore, the differences among various object classes are analyzed. In this way, this paper provides an in-depth analysis of the graph based object duplicate detection method.