TokenFormer: Rethinking Transformer Scaling with Tokenized Model Parameters

Transformers have become the predominant architecture in foundation models due to their excellent performance across various domains. However, the substantial cost of scaling these models remains a significant concern. This problem arises primarily from their dependence on a fixed number of parameters within linear projections. When architectural modifications (e.g., channel dimensions) are introduced, the entire model typically requires retraining from scratch. As model sizes continue growing, this strategy results in increasingly high computational costs and becomes unsustainable. To overcome this problem, we introduce TokenFormer, a natively scalable architecture that leverages the attention mechanism not only for computations among input tokens but also for interactions between tokens and model parameters, thereby enhancing architectural flexibility. By treating model parameters as tokens, we replace all the linear projections in Transformers with our token-parameter attention layer, where input tokens act as queries and model parameters as keys and values. This reformulation allows for progressive and efficient scaling without necessitating retraining from scratch. Our model scales from 124M to 1.4B parameters by incrementally adding new key-value parameter pairs, achieving performance comparable to Transformers trained from scratch while greatly reducing training costs. Code and models are available at \url{https://github.com/Haiyang-W/TokenFormer}.

Toward a Diffusion-Based Generalist for Dense Vision Tasks

Building generalized models that can solve many computer vision tasks simultaneously is an intriguing direction. Recent works have shown image itself can be used as a natural interface for general-purpose visual perception and demonstrated inspiring results. In this paper, we explore diffusion-based vision generalists, where we unify different types of dense prediction tasks as conditional image generation and re-purpose pre-trained diffusion models for it. However, directly applying off-the-shelf latent diffusion models leads to a quantization issue. Thus, we propose to perform diffusion in pixel space and provide a recipe for finetuning pre-trained text-to-image diffusion models for dense vision tasks. In experiments, we evaluate our method on four different types of tasks and show competitive performance to the other vision generalists.

LocCa: Visual Pretraining with Location-aware Captioners

Image captioning has been shown as an effective pretraining method similar to contrastive pretraining. However, the incorporation of location-aware information into visual pretraining remains an area with limited research. In this paper, we propose a simple visual pretraining method with location-aware captioners (LocCa). LocCa uses a simple image captioner task interface, to teach a model to read out rich information, i.e. bounding box coordinates, and captions, conditioned on the image pixel input. Thanks to the multitask capabilities of an encoder-decoder architecture, we show that an image captioner can easily handle multiple tasks during pretraining. Our experiments demonstrate that LocCa outperforms standard captioners significantly on localization downstream tasks while maintaining comparable performance on holistic tasks.

Text-Conditioned Resampler For Long Form Video Understanding

Videos are highly redundant data source and it is often enough to identify a few key moments to solve any given task. In this paper, we present a text-conditioned video resampler (TCR) module that uses a pre-trained and frozen visual encoder and large language model (LLM) to process long video sequences for a task. TCR localises relevant visual features from the video given a text condition and provides them to a LLM to generate a text response. Due to its lightweight design and use of cross-attention, TCR can process more than 100 frames at a time allowing the model to use much longer chunks of video than earlier works. We make the following contributions: (i) we design a transformer-based sampling architecture that can process long videos conditioned on a task, together with a training method that enables it to bridge pre-trained visual and language models; (ii) we empirically validate its efficacy on a wide variety of evaluation tasks, and set a new state-of-the-art on NextQA, EgoSchema, and the EGO4D-LTA challenge; and (iii) we determine tasks which require longer video contexts and that can thus be used effectively for further evaluation of long-range video models.

LIME: Localized Image Editing via Attention Regularization in Diffusion Models

Diffusion models (DMs) have gained prominence due to their ability to generate high-quality, varied images, with recent advancements in text-to-image generation. The research focus is now shifting towards the controllability of DMs. A significant challenge within this domain is localized editing, where specific areas of an image are modified without affecting the rest of the content. This paper introduces LIME for localized image editing in diffusion models that do not require user-specified regions of interest (RoI) or additional text input. Our method employs features from pre-trained methods and a simple clustering technique to obtain precise semantic segmentation maps. Then, by leveraging cross-attention maps, it refines these segments for localized edits. Finally, we propose a novel cross-attention regularization technique that penalizes unrelated cross-attention scores in the RoI during the denoising steps, ensuring localized edits. Our approach, without re-training and fine-tuning, consistently improves the performance of existing methods in various editing benchmarks.

LALM: Long-Term Action Anticipation with Language Models

Understanding human activity is a crucial yet intricate task in egocentric vision, a field that focuses on capturing visual perspectives from the camera wearer's viewpoint. While traditional methods heavily rely on representation learning trained on extensive video data, there exists a significant limitation: obtaining effective video representations proves challenging due to the inherent complexity and variability in human activities.Furthermore, exclusive dependence on video-based learning may constrain a model's capability to generalize across long-tail classes and out-of-distribution scenarios. In this study, we introduce a novel approach for long-term action anticipation using language models (LALM), adept at addressing the complex challenges of long-term activity understanding without the need for extensive training. Our method incorporates an action recognition model to track previous action sequences and a vision-language model to articulate relevant environmental details. By leveraging the context provided by these past events, we devise a prompting strategy for action anticipation using large language models (LLMs). Moreover, we implement Maximal Marginal Relevance for example selection to facilitate in-context learning of the LLMs. Our experimental results demonstrate that LALM surpasses the state-of-the-art methods in the task of long-term action anticipation on the Ego4D benchmark. We further validate LALM on two additional benchmarks, affirming its capacity for generalization across intricate activities with different sets of taxonomies. These are achieved without specific fine-tuning.

SILC: Improving Vision Language Pretraining with Self-Distillation

Image-Text pretraining on web-scale image caption dataset has become the default recipe for open vocabulary classification and retrieval models thanks to the success of CLIP and its variants. Several works have also used CLIP features for dense prediction tasks and have shown the emergence of open-set abilities. However, the contrastive objective only focuses on image-text alignment and does not incentivise image feature learning for dense prediction tasks. In this work, we propose the simple addition of local-to-global correspondence learning by self-distillation as an additional objective for contrastive pre-training to propose SILC. We show that distilling local image features from an exponential moving average (EMA) teacher model significantly improves model performance on several computer vision tasks including classification, retrieval, and especially segmentation. We further show that SILC scales better with the same training duration compared to the baselines. Our model SILC sets a new state of the art for zero-shot classification, few shot classification, image and text retrieval, zero-shot segmentation, and open vocabulary segmentation.

Detecting Adversarial Faces Using Only Real Face Self-Perturbations

Adversarial attacks aim to disturb the functionality of a target system by adding specific noise to the input samples, bringing potential threats to security and robustness when applied to facial recognition systems. Although existing defense techniques achieve high accuracy in detecting some specific adversarial faces (adv-faces), new attack methods especially GAN-based attacks with completely different noise patterns circumvent them and reach a higher attack success rate. Even worse, existing techniques require attack data before implementing the defense, making it impractical to defend newly emerging attacks that are unseen to defenders. In this paper, we investigate the intrinsic generality of adv-faces and propose to generate pseudo adv-faces by perturbing real faces with three heuristically designed noise patterns. We are the first to train an adv-face detector using only real faces and their self-perturbations, agnostic to victim facial recognition systems, and agnostic to unseen attacks. By regarding adv-faces as out-of-distribution data, we then naturally introduce a novel cascaded system for adv-face detection, which consists of training data self-perturbations, decision boundary regularization, and a max-pooling-based binary classifier focusing on abnormal local color aberrations. Experiments conducted on LFW and CelebA-HQ datasets with eight gradient-based and two GAN-based attacks validate that our method generalizes to a variety of unseen adversarial attacks.

Learning Prototype Classifiers for Long-Tailed Recognition

The problem of long-tailed recognition (LTR) has received attention in recent years due to the fundamental power-law distribution of objects in the real-world. Most recent works in LTR use softmax classifiers that have a tendency to correlate classifier norm with the amount of training data for a given class. On the other hand, Prototype classifiers do not suffer from this shortcoming and can deliver promising results simply using Nearest-Class-Mean (NCM), a special case where prototypes are empirical centroids. However, the potential of Prototype classifiers as an alternative to softmax in LTR is relatively underexplored. In this work, we propose Prototype classifiers, which jointly learn prototypes that minimize average cross-entropy loss based on probability scores from distances to prototypes. We theoretically analyze the properties of Euclidean distance based prototype classifiers that leads to stable gradient-based optimization which is robust to outliers. We further enhance Prototype classifiers by learning channel-dependent temperature parameters to enable independent distance scales along each channel. Our analysis shows that prototypes learned by Prototype classifiers are better separated than empirical centroids. Results on four long-tailed recognition benchmarks show that Prototype classifier outperforms or is comparable to the state-of-the-art methods.

Urban Scene Semantic Segmentation with Low-Cost Coarse Annotation

For best performance, today's semantic segmentation methods use large and carefully labeled datasets, requiring expensive annotation budgets. In this work, we show that coarse annotation is a low-cost but highly effective alternative for training semantic segmentation models. Considering the urban scene segmentation scenario, we leverage cheap coarse annotations for real-world captured data, as well as synthetic data to train our model and show competitive performance compared with finely annotated real-world data. Specifically, we propose a coarse-to-fine self-training framework that generates pseudo labels for unlabeled regions of the coarsely annotated data, using synthetic data to improve predictions around the boundaries between semantic classes, and using cross-domain data augmentation to increase diversity. Our extensive experimental results on Cityscapes and BDD100k datasets demonstrate that our method achieves a significantly better performance vs annotation cost tradeoff, yielding a comparable performance to fully annotated data with only a small fraction of the annotation budget. Also, when used as pretraining, our framework performs better compared to the standard fully supervised setting.