Ishan Misra

Maxime Oquab, Timothée Darcet, Théo Moutakanni et al. · meta-ai, mit

The recent breakthroughs in natural language processing for model pretraining on large quantities of data have opened the way for similar foundation models in computer vision. These models could greatly simplify the use of images in any system by producing all-purpose visual features, i.e., features that work across image distributions and tasks without finetuning. This work shows that existing pretraining methods, especially self-supervised methods, can produce such features if trained on enough curated data from diverse sources. We revisit existing approaches and combine different techniques to scale our pretraining in terms of data and model size. Most of the technical contributions aim at accelerating and stabilizing the training at scale. In terms of data, we propose an automatic pipeline to build a dedicated, diverse, and curated image dataset instead of uncurated data, as typically done in the self-supervised literature. In terms of models, we train a ViT model (Dosovitskiy et al., 2020) with 1B parameters and distill it into a series of smaller models that surpass the best available all-purpose features, OpenCLIP (Ilharco et al., 2021) on most of the benchmarks at image and pixel levels.

Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri et al. · allen-ai, berkeley

Modern artificial intelligence (AI) systems are powered by foundation models. This paper presents a new set of foundation models, called Llama 3. It is a herd of language models that natively support multilinguality, coding, reasoning, and tool usage. Our largest model is a dense Transformer with 405B parameters and a context window of up to 128K tokens. This paper presents an extensive empirical evaluation of Llama 3. We find that Llama 3 delivers comparable quality to leading language models such as GPT-4 on a plethora of tasks. We publicly release Llama 3, including pre-trained and post-trained versions of the 405B parameter language model and our Llama Guard 3 model for input and output safety. The paper also presents the results of experiments in which we integrate image, video, and speech capabilities into Llama 3 via a compositional approach. We observe this approach performs competitively with the state-of-the-art on image, video, and speech recognition tasks. The resulting models are not yet being broadly released as they are still under development.

Mahmoud Assran, Mathilde Caron, Ishan Misra et al. · meta-ai, mila

We propose Masked Siamese Networks (MSN), a self-supervised learning framework for learning image representations. Our approach matches the representation of an image view containing randomly masked patches to the representation of the original unmasked image. This self-supervised pre-training strategy is particularly scalable when applied to Vision Transformers since only the unmasked patches are processed by the network. As a result, MSNs improve the scalability of joint-embedding architectures, while producing representations of a high semantic level that perform competitively on low-shot image classification. For instance, on ImageNet-1K, with only 5,000 annotated images, our base MSN model achieves 72.4% top-1 accuracy, and with 1% of ImageNet-1K labels, we achieve 75.7% top-1 accuracy, setting a new state-of-the-art for self-supervised learning on this benchmark. Our code is publicly available.

Rohit Girdhar, Alaaeldin El-Nouby, Mannat Singh et al. · meta-ai

Transformer-based architectures have become competitive across a variety of visual domains, most notably images and videos. While prior work studies these modalities in isolation, having a common architecture suggests that one can train a single unified model for multiple visual modalities. Prior attempts at unified modeling typically use architectures tailored for vision tasks, or obtain worse performance compared to single modality models. In this work, we show that masked autoencoding can be used to train a simple Vision Transformer on images and videos, without requiring any labeled data. This single model learns visual representations that are comparable to or better than single-modality representations on both image and video benchmarks, while using a much simpler architecture. Furthermore, this model can be learned by dropping 90% of the image and 95% of the video patches, enabling extremely fast training of huge model architectures. In particular, we show that our single ViT-Huge model can be finetuned to achieve 86.6% on ImageNet and 75.5% on the challenging Something Something-v2 video benchmark, setting a new state-of-the-art.

Mannat Singh, Quentin Duval, Kalyan Vasudev Alwala et al. · meta-ai

This paper revisits the standard pretrain-then-finetune paradigm used in computer vision for visual recognition tasks. Typically, state-of-the-art foundation models are pretrained using large scale (weakly) supervised datasets with billions of images. We introduce an additional pre-pretraining stage that is simple and uses the self-supervised MAE technique to initialize the model. While MAE has only been shown to scale with the size of models, we find that it scales with the size of the training dataset as well. Thus, our MAE-based pre-pretraining scales with both model and data size making it applicable for training foundation models. Pre-pretraining consistently improves both the model convergence and the downstream transfer performance across a range of model scales (millions to billions of parameters), and dataset sizes (millions to billions of images). We measure the effectiveness of pre-pretraining on 10 different visual recognition tasks spanning image classification, video recognition, object detection, low-shot classification and zero-shot recognition. Our largest model achieves new state-of-the-art results on iNaturalist-18 (91.7%), ImageNet-ReaL (91.1%), 1-shot ImageNet-1k (63.6%), and zero-shot transfer on Food-101 (96.2%). Our study reveals that model initialization plays a significant role, even for web-scale pretraining with billions of images, and our models are available publicly.

Xudong Wang, Rohit Girdhar, Stella X. Yu et al. · meta-ai

We propose Cut-and-LEaRn (CutLER), a simple approach for training unsupervised object detection and segmentation models. We leverage the property of self-supervised models to 'discover' objects without supervision and amplify it to train a state-of-the-art localization model without any human labels. CutLER first uses our proposed MaskCut approach to generate coarse masks for multiple objects in an image and then learns a detector on these masks using our robust loss function. We further improve the performance by self-training the model on its predictions. Compared to prior work, CutLER is simpler, compatible with different detection architectures, and detects multiple objects. CutLER is also a zero-shot unsupervised detector and improves detection performance AP50 by over 2.7 times on 11 benchmarks across domains like video frames, paintings, sketches, etc. With finetuning, CutLER serves as a low-shot detector surpassing MoCo-v2 by 7.3% APbox and 6.6% APmask on COCO when training with 5% labels.

Xudong Wang, Ishan Misra, Ziyun Zeng et al. · meta-ai

Existing approaches to unsupervised video instance segmentation typically rely on motion estimates and experience difficulties tracking small or divergent motions. We present VideoCutLER, a simple method for unsupervised multi-instance video segmentation without using motion-based learning signals like optical flow or training on natural videos. Our key insight is that using high-quality pseudo masks and a simple video synthesis method for model training is surprisingly sufficient to enable the resulting video model to effectively segment and track multiple instances across video frames. We show the first competitive unsupervised learning results on the challenging YouTubeVIS-2019 benchmark, achieving 50.7% APvideo^50 , surpassing the previous state-of-the-art by a large margin. VideoCutLER can also serve as a strong pretrained model for supervised video instance segmentation tasks, exceeding DINO by 15.9% on YouTubeVIS-2019 in terms of APvideo.

Yue Zhao, Ishan Misra, Philipp Krähenbühl et al. · meta-ai

We introduce LaViLa, a new approach to learning video-language representations by leveraging Large Language Models (LLMs). We repurpose pre-trained LLMs to be conditioned on visual input, and finetune them to create automatic video narrators. Our auto-generated narrations offer a number of advantages, including dense coverage of long videos, better temporal synchronization of the visual information and text, and much higher diversity of text. The video-text embedding learned contrastively with these additional auto-generated narrations outperforms the previous state-of-the-art on multiple first-person and third-person video tasks, both in zero-shot and finetuned setups. Most notably, LaViLa obtains an absolute gain of 10.1% on EGTEA classification and 5.9% Epic-Kitchens-100 multi-instance retrieval benchmarks. Furthermore, LaViLa trained with only half the narrations from the Ego4D dataset outperforms baseline models trained on the full set, and shows positive scaling behavior on increasing pre-training data and model size.

Rohit Girdhar, Mannat Singh, Andrew Brown et al. · meta-ai

We present Emu Video, a text-to-video generation model that factorizes the generation into two steps: first generating an image conditioned on the text, and then generating a video conditioned on the text and the generated image. We identify critical design decisions--adjusted noise schedules for diffusion, and multi-stage training that enable us to directly generate high quality and high resolution videos, without requiring a deep cascade of models as in prior work. In human evaluations, our generated videos are strongly preferred in quality compared to all prior work--81% vs. Google's Imagen Video, 90% vs. Nvidia's PYOCO, and 96% vs. Meta's Make-A-Video. Our model outperforms commercial solutions such as RunwayML's Gen2 and Pika Labs. Finally, our factorizing approach naturally lends itself to animating images based on a user's text prompt, where our generations are preferred 96% over prior work.

Nur Muhammad Mahi Shafiullah, Anant Rai, Haritheja Etukuru et al.

Throughout history, we have successfully integrated various machines into our homes. Dishwashers, laundry machines, stand mixers, and robot vacuums are a few recent examples. However, these machines excel at performing only a single task effectively. The concept of a "generalist machine" in homes - a domestic assistant that can adapt and learn from our needs, all while remaining cost-effective - has long been a goal in robotics that has been steadily pursued for decades. In this work, we initiate a large-scale effort towards this goal by introducing Dobb-E, an affordable yet versatile general-purpose system for learning robotic manipulation within household settings. Dobb-E can learn a new task with only five minutes of a user showing it how to do it, thanks to a demonstration collection tool ("The Stick") we built out of cheap parts and iPhones. We use the Stick to collect 13 hours of data in 22 homes of New York City, and train Home Pretrained Representations (HPR). Then, in a novel home environment, with five minutes of demonstrations and fifteen minutes of adapting the HPR model, we show that Dobb-E can reliably solve the task on the Stretch, a mobile robot readily available on the market. Across roughly 30 days of experimentation in homes of New York City and surrounding areas, we test our system in 10 homes, with a total of 109 tasks in different environments, and finally achieve a success rate of 81%. Beyond success percentages, our experiments reveal a plethora of unique challenges absent or ignored in lab robotics. These range from effects of strong shadows, to variable demonstration quality by non-expert users. With the hope of accelerating research on home robots, and eventually seeing robot butlers in every home, we open-source Dobb-E software stack and models, our data, and our hardware designs at https://dobb-e.com

Mahmoud Assran, Quentin Duval, Ishan Misra et al.

This paper demonstrates an approach for learning highly semantic image representations without relying on hand-crafted data-augmentations. We introduce the Image-based Joint-Embedding Predictive Architecture (I-JEPA), a non-generative approach for self-supervised learning from images. The idea behind I-JEPA is simple: from a single context block, predict the representations of various target blocks in the same image. A core design choice to guide I-JEPA towards producing semantic representations is the masking strategy; specifically, it is crucial to (a) sample target blocks with sufficiently large scale (semantic), and to (b) use a sufficiently informative (spatially distributed) context block. Empirically, when combined with Vision Transformers, we find I-JEPA to be highly scalable. For instance, we train a ViT-Huge/14 on ImageNet using 16 A100 GPUs in under 72 hours to achieve strong downstream performance across a wide range of tasks, from linear classification to object counting and depth prediction.

Yang Fu, Ishan Misra, Xiaolong Wang · ibm-research

We propose a generalizable neural radiance fields - MonoNeRF, that can be trained on large-scale monocular videos of moving in static scenes without any ground-truth annotations of depth and camera poses. MonoNeRF follows an Autoencoder-based architecture, where the encoder estimates the monocular depth and the camera pose, and the decoder constructs a Multiplane NeRF representation based on the depth encoder feature, and renders the input frames with the estimated camera. The learning is supervised by the reconstruction error. Once the model is learned, it can be applied to multiple applications including depth estimation, camera pose estimation, and single-image novel view synthesis. More qualitative results are available at: https://oasisyang.github.io/mononerf .

Sai Saketh Rambhatla, Ishan Misra, Rama Chellappa et al.

We tackle the challenging task of unsupervised object localization in this work. Recently, transformers trained with self-supervised learning have been shown to exhibit object localization properties without being trained for this task. In this work, we present Multiple Object localization with Self-supervised Transformers (MOST) that uses features of transformers trained using self-supervised learning to localize multiple objects in real world images. MOST analyzes the similarity maps of the features using box counting; a fractal analysis tool to identify tokens lying on foreground patches. The identified tokens are then clustered together, and tokens of each cluster are used to generate bounding boxes on foreground regions. Unlike recent state-of-the-art object localization methods, MOST can localize multiple objects per image and outperforms SOTA algorithms on several object localization and discovery benchmarks on PASCAL-VOC 07, 12 and COCO20k datasets. Additionally, we show that MOST can be used for self-supervised pre-training of object detectors, and yields consistent improvements on fully, semi-supervised object detection and unsupervised region proposal generation.

Mahmoud Assran, Randall Balestriero, Quentin Duval et al.

A successful paradigm in representation learning is to perform self-supervised pretraining using tasks based on mini-batch statistics (e.g., SimCLR, VICReg, SwAV, MSN). We show that in the formulation of all these methods is an overlooked prior to learn features that enable uniform clustering of the data. While this prior has led to remarkably semantic representations when pretraining on class-balanced data, such as ImageNet, we demonstrate that it can hamper performance when pretraining on class-imbalanced data. By moving away from conventional uniformity priors and instead preferring power-law distributed feature clusters, we show that one can improve the quality of the learned representations on real-world class-imbalanced datasets. To demonstrate this, we develop an extension of the Masked Siamese Networks (MSN) method to support the use of arbitrary features priors.

Kumar Ashutosh, XuDong Wang, Xi Yin et al. · meta-ai

Video generation models have recently achieved impressive visual fidelity and temporal coherence. Yet, they continue to struggle with complex, non-rigid motions, especially when synthesizing humans performing dynamic actions such as sports, dance, etc. Generated videos often exhibit missing or extra limbs, distorted poses, or physically implausible actions. In this work, we propose a remarkably simple reward model, HuDA, to quantify and improve the human motion in generated videos. HuDA integrates human detection confidence for appearance quality, and a temporal prompt alignment score to capture motion realism. We show this simple reward function that leverages off-the-shelf models without any additional training, outperforms specialized models finetuned with manually annotated data. Using HuDA for Group Reward Policy Optimization (GRPO) post-training of video models, we significantly enhance video generation, especially when generating complex human motions, outperforming state-of-the-art models like Wan 2.1, with win-rate of 73%. Finally, we demonstrate that HuDA improves generation quality beyond just humans, for instance, significantly improving generation of animal videos and human-object interactions.

Sagar Vaze, Nicolas Carion, Ishan Misra

We argue that there are many notions of 'similarity' and that models, like humans, should be able to adapt to these dynamically. This contrasts with most representation learning methods, supervised or self-supervised, which learn a fixed embedding function and hence implicitly assume a single notion of similarity. For instance, models trained on ImageNet are biased towards object categories, while a user might prefer the model to focus on colors, textures or specific elements in the scene. In this paper, we propose the GeneCIS ('genesis') benchmark, which measures models' ability to adapt to a range of similarity conditions. Extending prior work, our benchmark is designed for zero-shot evaluation only, and hence considers an open-set of similarity conditions. We find that baselines from powerful CLIP models struggle on GeneCIS and that performance on the benchmark is only weakly correlated with ImageNet accuracy, suggesting that simply scaling existing methods is not fruitful. We further propose a simple, scalable solution based on automatically mining information from existing image-caption datasets. We find our method offers a substantial boost over the baselines on GeneCIS, and further improves zero-shot performance on related image retrieval benchmarks. In fact, though evaluated zero-shot, our model surpasses state-of-the-art supervised models on MIT-States. Project page at https://sgvaze.github.io/genecis/.

Melissa Hall, Laura Gustafson, Aaron Adcock et al.

We explore the extent to which zero-shot vision-language models exhibit gender bias for different vision tasks. Vision models traditionally required task-specific labels for representing concepts, as well as finetuning; zero-shot models like CLIP instead perform tasks with an open-vocabulary, meaning they do not need a fixed set of labels, by using text embeddings to represent concepts. With these capabilities in mind, we ask: Do vision-language models exhibit gender bias when performing zero-shot image classification, object detection and semantic segmentation? We evaluate different vision-language models with multiple datasets across a set of concepts and find (i) all models evaluated show distinct performance differences based on the perceived gender of the person co-occurring with a given concept in the image and that aggregating analyses over all concepts can mask these concerns; (ii) model calibration (i.e. the relationship between accuracy and confidence) also differs distinctly by perceived gender, even when evaluating on similar representations of concepts; and (iii) these observed disparities align with existing gender biases in word embeddings from language models. These findings suggest that, while language greatly expands the capability of vision tasks, it can also contribute to social biases in zero-shot vision settings. Furthermore, biases can further propagate when foundational models like CLIP are used by other models to enable zero-shot capabilities.

Sangwoo Mo, Jong-Chyi Su, Chih-Yao Ma et al.

Semi-supervised learning aims to train a model using limited labels. State-of-the-art semi-supervised methods for image classification such as PAWS rely on self-supervised representations learned with large-scale unlabeled but curated data. However, PAWS is often less effective when using real-world unlabeled data that is uncurated, e.g., contains out-of-class data. We propose RoPAWS, a robust extension of PAWS that can work with real-world unlabeled data. We first reinterpret PAWS as a generative classifier that models densities using kernel density estimation. From this probabilistic perspective, we calibrate its prediction based on the densities of labeled and unlabeled data, which leads to a simple closed-form solution from the Bayes' rule. We demonstrate that RoPAWS significantly improves PAWS for uncurated Semi-iNat by +5.3% and curated ImageNet by +0.4%.

Quentin Duval, Ishan Misra, Nicolas Ballas

Self-supervised methods in vision have been mostly focused on large architectures as they seem to suffer from a significant performance drop for smaller architectures. In this paper, we propose a simple self-supervised distillation technique that can train high performance low-compute neural networks. Our main insight is that existing joint-embedding based SSL methods can be repurposed for knowledge distillation from a large self-supervised teacher to a small student model. Thus, we call our method Replace one Branch (RoB) as it simply replaces one branch of the joint-embedding training with a large teacher model. RoB is widely applicable to a number of architectures such as small ResNets, MobileNets and ViT, and pretrained models such as DINO, SwAV or iBOT. When pretraining on the ImageNet dataset, RoB yields models that compete with supervised knowledge distillation. When applied to MSN, RoB produces students with strong semi-supervised capabilities. Finally, our best ViT-Tiny models improve over prior SSL state-of-the-art on ImageNet by $2.3\%$ and are on par or better than a supervised distilled DeiT on five downstream transfer tasks (iNaturalist, CIFAR, Clevr/Count, Clevr/Dist and Places). We hope RoB enables practical self-supervision at smaller scale.

Sai Saketh Rambhatla, Ishan Misra

We present an automated way to evaluate the text alignment of text-to-image generative diffusion models using standard image-text recognition datasets. Our method, called SelfEval, uses the generative model to compute the likelihood of real images given text prompts, and the likelihood can be used to perform recognition tasks with the generative model. We evaluate generative models on standard datasets created for multimodal text-image discriminative learning and assess fine-grained aspects of their performance: attribute binding, color recognition, counting, shape recognition, spatial understanding. Existing automated metrics rely on an external pretrained model like CLIP (VLMs) or LLMs, and are sensitive to the exact pretrained model and its limitations. SelfEval sidesteps these issues, and to the best of our knowledge, is the first automated metric to show a high degree of agreement for measuring text-faithfulness with the gold-standard human evaluations across multiple generative models, benchmarks and evaluation metrics. SelfEval also reveals that generative models showcase competitive recognition performance on challenging tasks such as Winoground image-score compared to discriminative models. We hope SelfEval enables easy and reliable automated evaluation for diffusion models.

Adam Polyak, Amit Zohar, Andrew Brown et al. · meta-ai

We present Movie Gen, a cast of foundation models that generates high-quality, 1080p HD videos with different aspect ratios and synchronized audio. We also show additional capabilities such as precise instruction-based video editing and generation of personalized videos based on a user's image. Our models set a new state-of-the-art on multiple tasks: text-to-video synthesis, video personalization, video editing, video-to-audio generation, and text-to-audio generation. Our largest video generation model is a 30B parameter transformer trained with a maximum context length of 73K video tokens, corresponding to a generated video of 16 seconds at 16 frames-per-second. We show multiple technical innovations and simplifications on the architecture, latent spaces, training objectives and recipes, data curation, evaluation protocols, parallelization techniques, and inference optimizations that allow us to reap the benefits of scaling pre-training data, model size, and training compute for training large scale media generation models. We hope this paper helps the research community to accelerate progress and innovation in media generation models. All videos from this paper are available at https://go.fb.me/MovieGenResearchVideos.

Xudong Wang, Trevor Darrell, Sai Saketh Rambhatla et al. · meta-ai

Text-to-image diffusion models produce high quality images but do not offer control over individual instances in the image. We introduce InstanceDiffusion that adds precise instance-level control to text-to-image diffusion models. InstanceDiffusion supports free-form language conditions per instance and allows flexible ways to specify instance locations such as simple single points, scribbles, bounding boxes or intricate instance segmentation masks, and combinations thereof. We propose three major changes to text-to-image models that enable precise instance-level control. Our UniFusion block enables instance-level conditions for text-to-image models, the ScaleU block improves image fidelity, and our Multi-instance Sampler improves generations for multiple instances. InstanceDiffusion significantly surpasses specialized state-of-the-art models for each location condition. Notably, on the COCO dataset, we outperform previous state-of-the-art by 20.4% AP$_{50}^\text{box}$ for box inputs, and 25.4% IoU for mask inputs.

Xingyi Zhou, Rohit Girdhar, Armand Joulin et al.

Current object detectors are limited in vocabulary size due to the small scale of detection datasets. Image classifiers, on the other hand, reason about much larger vocabularies, as their datasets are larger and easier to collect. We propose Detic, which simply trains the classifiers of a detector on image classification data and thus expands the vocabulary of detectors to tens of thousands of concepts. Unlike prior work, Detic does not need complex assignment schemes to assign image labels to boxes based on model predictions, making it much easier to implement and compatible with a range of detection architectures and backbones. Our results show that Detic yields excellent detectors even for classes without box annotations. It outperforms prior work on both open-vocabulary and long-tail detection benchmarks. Detic provides a gain of 2.4 mAP for all classes and 8.3 mAP for novel classes on the open-vocabulary LVIS benchmark. On the standard LVIS benchmark, Detic obtains 41.7 mAP when evaluated on all classes, or only rare classes, hence closing the gap in performance for object categories with few samples. For the first time, we train a detector with all the twenty-one-thousand classes of the ImageNet dataset and show that it generalizes to new datasets without finetuning. Code is available at \url{https://github.com/facebookresearch/Detic}.

Mandela Patrick, Dylan Campbell, Yuki M. Asano et al.

In video transformers, the time dimension is often treated in the same way as the two spatial dimensions. However, in a scene where objects or the camera may move, a physical point imaged at one location in frame $t$ may be entirely unrelated to what is found at that location in frame $t+k$. These temporal correspondences should be modeled to facilitate learning about dynamic scenes. To this end, we propose a new drop-in block for video transformers -- trajectory attention -- that aggregates information along implicitly determined motion paths. We additionally propose a new method to address the quadratic dependence of computation and memory on the input size, which is particularly important for high resolution or long videos. While these ideas are useful in a range of settings, we apply them to the specific task of video action recognition with a transformer model and obtain state-of-the-art results on the Kinetics, Something--Something V2, and Epic-Kitchens datasets. Code and models are available at: https://github.com/facebookresearch/Motionformer

Aishwarya Kamath, Mannat Singh, Yann LeCun et al.

Multi-modal reasoning systems rely on a pre-trained object detector to extract regions of interest from the image. However, this crucial module is typically used as a black box, trained independently of the downstream task and on a fixed vocabulary of objects and attributes. This makes it challenging for such systems to capture the long tail of visual concepts expressed in free form text. In this paper we propose MDETR, an end-to-end modulated detector that detects objects in an image conditioned on a raw text query, like a caption or a question. We use a transformer-based architecture to reason jointly over text and image by fusing the two modalities at an early stage of the model. We pre-train the network on 1.3M text-image pairs, mined from pre-existing multi-modal datasets having explicit alignment between phrases in text and objects in the image. We then fine-tune on several downstream tasks such as phrase grounding, referring expression comprehension and segmentation, achieving state-of-the-art results on popular benchmarks. We also investigate the utility of our model as an object detector on a given label set when fine-tuned in a few-shot setting. We show that our pre-training approach provides a way to handle the long tail of object categories which have very few labelled instances. Our approach can be easily extended for visual question answering, achieving competitive performance on GQA and CLEVR. The code and models are available at https://github.com/ashkamath/mdetr.

Priya Goyal, Mathilde Caron, Benjamin Lefaudeux et al.

Recently, self-supervised learning methods like MoCo, SimCLR, BYOL and SwAV have reduced the gap with supervised methods. These results have been achieved in a control environment, that is the highly curated ImageNet dataset. However, the premise of self-supervised learning is that it can learn from any random image and from any unbounded dataset. In this work, we explore if self-supervision lives to its expectation by training large models on random, uncurated images with no supervision. Our final SElf-supERvised (SEER) model, a RegNetY with 1.3B parameters trained on 1B random images with 512 GPUs achieves 84.2% top-1 accuracy, surpassing the best self-supervised pretrained model by 1% and confirming that self-supervised learning works in a real world setting. Interestingly, we also observe that self-supervised models are good few-shot learners achieving 77.9% top-1 with access to only 10% of ImageNet. Code: https://github.com/facebookresearch/vissl

Priya Goyal, Dhruv Mahajan, Abhinav Gupta et al.

Self-supervised learning aims to learn representations from the data itself without explicit manual supervision. Existing efforts ignore a crucial aspect of self-supervised learning - the ability to scale to large amount of data because self-supervision requires no manual labels. In this work, we revisit this principle and scale two popular self-supervised approaches to 100 million images. We show that by scaling on various axes (including data size and problem 'hardness'), one can largely match or even exceed the performance of supervised pre-training on a variety of tasks such as object detection, surface normal estimation (3D) and visual navigation using reinforcement learning. Scaling these methods also provides many interesting insights into the limitations of current self-supervised techniques and evaluations. We conclude that current self-supervised methods are not 'hard' enough to take full advantage of large scale data and do not seem to learn effective high level semantic representations. We also introduce an extensive benchmark across 9 different datasets and tasks. We believe that such a benchmark along with comparable evaluation settings is necessary to make meaningful progress. Code is at: https://github.com/facebookresearch/fair_self_supervision_benchmark.

Feng Liang, Bichen Wu, Jialiang Wang et al.

Diffusion models have transformed the image-to-image (I2I) synthesis and are now permeating into videos. However, the advancement of video-to-video (V2V) synthesis has been hampered by the challenge of maintaining temporal consistency across video frames. This paper proposes a consistent V2V synthesis framework by jointly leveraging spatial conditions and temporal optical flow clues within the source video. Contrary to prior methods that strictly adhere to optical flow, our approach harnesses its benefits while handling the imperfection in flow estimation. We encode the optical flow via warping from the first frame and serve it as a supplementary reference in the diffusion model. This enables our model for video synthesis by editing the first frame with any prevalent I2I models and then propagating edits to successive frames. Our V2V model, FlowVid, demonstrates remarkable properties: (1) Flexibility: FlowVid works seamlessly with existing I2I models, facilitating various modifications, including stylization, object swaps, and local edits. (2) Efficiency: Generation of a 4-second video with 30 FPS and 512x512 resolution takes only 1.5 minutes, which is 3.1x, 7.2x, and 10.5x faster than CoDeF, Rerender, and TokenFlow, respectively. (3) High-quality: In user studies, our FlowVid is preferred 45.7% of the time, outperforming CoDeF (3.5%), Rerender (10.2%), and TokenFlow (40.4%).

Yinbo Chen, Rohit Girdhar, Xiaolong Wang et al. · meta-ai

Tokenizing images into compact visual representations is a key step in learning efficient and high-quality image generative models. We present a simple diffusion tokenizer (DiTo) that learns compact visual representations for image generation models. Our key insight is that a single learning objective, diffusion L2 loss, can be used for training scalable image tokenizers. Since diffusion is already widely used for image generation, our insight greatly simplifies training such tokenizers. In contrast, current state-of-the-art tokenizers rely on an empirically found combination of heuristics and losses, thus requiring a complex training recipe that relies on non-trivially balancing different losses and pretrained supervised models. We show design decisions, along with theoretical grounding, that enable us to scale DiTo for learning competitive image representations. Our results show that DiTo is a simpler, scalable, and self-supervised alternative to the current state-of-the-art image tokenizer which is supervised. DiTo achieves competitive or better quality than state-of-the-art in image reconstruction and downstream image generation tasks.

Nupur Kumari, Xi Yin, Jun-Yan Zhu et al.

Customization of text-to-image models enables users to insert new concepts or objects and generate them in unseen settings. Existing methods either rely on comparatively expensive test-time optimization or train encoders on single-image datasets without multi-image supervision, which can limit image quality. We propose a simple approach to address these challenges. We first leverage existing text-to-image models and 3D datasets to create a high-quality Synthetic Customization Dataset (SynCD) consisting of multiple images of the same object in different lighting, backgrounds, and poses. Using this dataset, we train an encoder-based model that incorporates fine-grained visual details from reference images via a shared attention mechanism. Finally, we propose an inference technique that normalizes text and image guidance vectors to mitigate overexposure issues in sampled images. Through extensive experiments, we show that our encoder-based model, trained on SynCD, and with the proposed inference algorithm, improves upon existing encoder-based methods on standard customization benchmarks.

Kumar Ashutosh, Yossi Gandelsman, Xinlei Chen et al. · meta-ai

We present MILS: Multimodal Iterative LLM Solver, a surprisingly simple, training-free approach, to imbue multimodal capabilities into your favorite LLM. Leveraging their innate ability to perform multi-step reasoning, MILS prompts the LLM to generate candidate outputs, each of which are scored and fed back iteratively, eventually generating a solution to the task. This enables various applications that typically require training specialized models on task-specific data. In particular, we establish a new state-of-the-art on emergent zero-shot image, video and audio captioning. MILS seamlessly applies to media generation as well, discovering prompt rewrites to improve text-to-image generation, and even edit prompts for style transfer! Finally, being a gradient-free optimization approach, MILS can invert multimodal embeddings into text, enabling applications like cross-modal arithmetic.

Sachit Menon, Ishan Misra, Rohit Girdhar · meta-ai

We introduce the new task of generating Illustrated Instructions, i.e., visual instructions customized to a user's needs. We identify desiderata unique to this task, and formalize it through a suite of automatic and human evaluation metrics, designed to measure the validity, consistency, and efficacy of the generations. We combine the power of large language models (LLMs) together with strong text-to-image generation diffusion models to propose a simple approach called StackedDiffusion, which generates such illustrated instructions given text as input. The resulting model strongly outperforms baseline approaches and state-of-the-art multimodal LLMs; and in 30% of cases, users even prefer it to human-generated articles. Most notably, it enables various new and exciting applications far beyond what static articles on the web can provide, such as personalized instructions complete with intermediate steps and pictures in response to a user's individual situation.

Junhong Shen, Kushal Tirumala, Michihiro Yasunaga et al.

Most existing image tokenizers encode images into a fixed number of tokens or patches, overlooking the inherent variability in image complexity. To address this, we introduce Content-Adaptive Tokenizer (CAT), which dynamically adjusts representation capacity based on the image content and encodes simpler images into fewer tokens. We design a caption-based evaluation system that leverages large language models (LLMs) to predict content complexity and determine the optimal compression ratio for a given image, taking into account factors critical to human perception. Trained on images with diverse compression ratios, CAT demonstrates robust performance in image reconstruction. We also utilize its variable-length latent representations to train Diffusion Transformers (DiTs) for ImageNet generation. By optimizing token allocation, CAT improves the FID score over fixed-ratio baselines trained with the same flops and boosts the inference throughput by 18.5%.

Bolin Lai, Xudong Wang, Saketh Rambhatla et al.

Latent diffusion has become the default paradigm for visual generation, yet we observe a persistent reconstruction-generation trade-off as latent dimensionality increases: higher-capacity autoencoders improve reconstruction fidelity but generation quality eventually declines. We trace this gap to the different behaviors in high-frequency encoding and decoding. Through controlled perturbations in both RGB and latent domains, we analyze encoder/decoder behaviors and find that decoders depend strongly on high-frequency latent components to recover details, whereas encoders under-represent high-frequency contents, yielding insufficient exposure and underfitting in high-frequency bands for diffusion model training. To address this issue, we introduce FreqWarm, a plug-and-play frequency warm-up curriculum that increases early-stage exposure to high-frequency latent signals during diffusion or flow-matching training -- without modifying or retraining the autoencoder. Applied across several high-dimensional autoencoders, FreqWarm consistently improves generation quality: decreasing gFID by 14.11 on Wan2.2-VAE, 6.13 on LTX-VAE, and 4.42 on DC-AE-f32, while remaining architecture-agnostic and compatible with diverse backbones. Our study shows that explicitly managing frequency exposure can successfully turn high-dimensional latent spaces into more diffusible targets.

Rohit Girdhar, Alaaeldin El-Nouby, Zhuang Liu et al.

We present ImageBind, an approach to learn a joint embedding across six different modalities - images, text, audio, depth, thermal, and IMU data. We show that all combinations of paired data are not necessary to train such a joint embedding, and only image-paired data is sufficient to bind the modalities together. ImageBind can leverage recent large scale vision-language models, and extends their zero-shot capabilities to new modalities just by using their natural pairing with images. It enables novel emergent applications 'out-of-the-box' including cross-modal retrieval, composing modalities with arithmetic, cross-modal detection and generation. The emergent capabilities improve with the strength of the image encoder and we set a new state-of-the-art on emergent zero-shot recognition tasks across modalities, outperforming specialist supervised models. Finally, we show strong few-shot recognition results outperforming prior work, and that ImageBind serves as a new way to evaluate vision models for visual and non-visual tasks.

Priya Goyal, Quentin Duval, Isaac Seessel et al.

Discriminative self-supervised learning allows training models on any random group of internet images, and possibly recover salient information that helps differentiate between the images. Applied to ImageNet, this leads to object centric features that perform on par with supervised features on most object-centric downstream tasks. In this work, we question if using this ability, we can learn any salient and more representative information present in diverse unbounded set of images from across the globe. To do so, we train models on billions of random images without any data pre-processing or prior assumptions about what we want the model to learn. We scale our model size to dense 10 billion parameters to avoid underfitting on a large data size. We extensively study and validate our model performance on over 50 benchmarks including fairness, robustness to distribution shift, geographical diversity, fine grained recognition, image copy detection and many image classification datasets. The resulting model, not only captures well semantic information, it also captures information about artistic style and learns salient information such as geolocations and multilingual word embeddings based on visual content only. More importantly, we discover that such model is more robust, more fair, less harmful and less biased than supervised models or models trained on object centric datasets such as ImageNet.

Randall Balestriero, Ishan Misra, Yann LeCun

Data-Augmentation (DA) is known to improve performance across tasks and datasets. We propose a method to theoretically analyze the effect of DA and study questions such as: how many augmented samples are needed to correctly estimate the information encoded by that DA? How does the augmentation policy impact the final parameters of a model? We derive several quantities in close-form, such as the expectation and variance of an image, loss, and model's output under a given DA distribution. Those derivations open new avenues to quantify the benefits and limitations of DA. For example, we show that common DAs require tens of thousands of samples for the loss at hand to be correctly estimated and for the model training to converge. We show that for a training loss to be stable under DA sampling, the model's saliency map (gradient of the loss with respect to the model's input) must align with the smallest eigenvector of the sample variance under the considered DA augmentation, hinting at a possible explanation on why models tend to shift their focus from edges to textures.

Rohit Girdhar, Mannat Singh, Nikhila Ravi et al.

Prior work has studied different visual modalities in isolation and developed separate architectures for recognition of images, videos, and 3D data. Instead, in this paper, we propose a single model which excels at classifying images, videos, and single-view 3D data using exactly the same model parameters. Our 'Omnivore' model leverages the flexibility of transformer-based architectures and is trained jointly on classification tasks from different modalities. Omnivore is simple to train, uses off-the-shelf standard datasets, and performs at-par or better than modality-specific models of the same size. A single Omnivore model obtains 86.0% on ImageNet, 84.1% on Kinetics, and 67.1% on SUN RGB-D. After finetuning, our models outperform prior work on a variety of vision tasks and generalize across modalities. Omnivore's shared visual representation naturally enables cross-modal recognition without access to correspondences between modalities. We hope our results motivate researchers to model visual modalities together.

Bowen Cheng, Anwesa Choudhuri, Ishan Misra et al.

We find Mask2Former also achieves state-of-the-art performance on video instance segmentation without modifying the architecture, the loss or even the training pipeline. In this report, we show universal image segmentation architectures trivially generalize to video segmentation by directly predicting 3D segmentation volumes. Specifically, Mask2Former sets a new state-of-the-art of 60.4 AP on YouTubeVIS-2019 and 52.6 AP on YouTubeVIS-2021. We believe Mask2Former is also capable of handling video semantic and panoptic segmentation, given its versatility in image segmentation. We hope this will make state-of-the-art video segmentation research more accessible and bring more attention to designing universal image and video segmentation architectures.

Bowen Cheng, Ishan Misra, Alexander G. Schwing et al.

Image segmentation is about grouping pixels with different semantics, e.g., category or instance membership, where each choice of semantics defines a task. While only the semantics of each task differ, current research focuses on designing specialized architectures for each task. We present Masked-attention Mask Transformer (Mask2Former), a new architecture capable of addressing any image segmentation task (panoptic, instance or semantic). Its key components include masked attention, which extracts localized features by constraining cross-attention within predicted mask regions. In addition to reducing the research effort by at least three times, it outperforms the best specialized architectures by a significant margin on four popular datasets. Most notably, Mask2Former sets a new state-of-the-art for panoptic segmentation (57.8 PQ on COCO), instance segmentation (50.1 AP on COCO) and semantic segmentation (57.7 mIoU on ADE20K).

Omri Puny, Matan Atzmon, Heli Ben-Hamu et al.

Many machine learning tasks involve learning functions that are known to be invariant or equivariant to certain symmetries of the input data. However, it is often challenging to design neural network architectures that respect these symmetries while being expressive and computationally efficient. For example, Euclidean motion invariant/equivariant graph or point cloud neural networks. We introduce Frame Averaging (FA), a general purpose and systematic framework for adapting known (backbone) architectures to become invariant or equivariant to new symmetry types. Our framework builds on the well known group averaging operator that guarantees invariance or equivariance but is intractable. In contrast, we observe that for many important classes of symmetries, this operator can be replaced with an averaging operator over a small subset of the group elements, called a frame. We show that averaging over a frame guarantees exact invariance or equivariance while often being much simpler to compute than averaging over the entire group. Furthermore, we prove that FA-based models have maximal expressive power in a broad setting and in general preserve the expressive power of their backbone architectures. Using frame averaging, we propose a new class of universal Graph Neural Networks (GNNs), universal Euclidean motion invariant point cloud networks, and Euclidean motion invariant Message Passing (MP) GNNs. We demonstrate the practical effectiveness of FA on several applications including point cloud normal estimation, beyond $2$-WL graph separation, and $n$-body dynamics prediction, achieving state-of-the-art results in all of these benchmarks.

Ishan Misra, Rohit Girdhar, Armand Joulin

We propose 3DETR, an end-to-end Transformer based object detection model for 3D point clouds. Compared to existing detection methods that employ a number of 3D-specific inductive biases, 3DETR requires minimal modifications to the vanilla Transformer block. Specifically, we find that a standard Transformer with non-parametric queries and Fourier positional embeddings is competitive with specialized architectures that employ libraries of 3D-specific operators with hand-tuned hyperparameters. Nevertheless, 3DETR is conceptually simple and easy to implement, enabling further improvements by incorporating 3D domain knowledge. Through extensive experiments, we show 3DETR outperforms the well-established and highly optimized VoteNet baselines on the challenging ScanNetV2 dataset by 9.5%. Furthermore, we show 3DETR is applicable to 3D tasks beyond detection, and can serve as a building block for future research.

Zhongzheng Ren, Ishan Misra, Alexander G. Schwing et al.

We introduce WyPR, a Weakly-supervised framework for Point cloud Recognition, requiring only scene-level class tags as supervision. WyPR jointly addresses three core 3D recognition tasks: point-level semantic segmentation, 3D proposal generation, and 3D object detection, coupling their predictions through self and cross-task consistency losses. We show that in conjunction with standard multiple-instance learning objectives, WyPR can detect and segment objects in point cloud data without access to any spatial labels at training time. We demonstrate its efficacy using the ScanNet and S3DIS datasets, outperforming prior state of the art on weakly-supervised segmentation by more than 6% mIoU. In addition, we set up the first benchmark for weakly-supervised 3D object detection on both datasets, where WyPR outperforms standard approaches and establishes strong baselines for future work.

Mathilde Caron, Hugo Touvron, Ishan Misra et al.

In this paper, we question if self-supervised learning provides new properties to Vision Transformer (ViT) that stand out compared to convolutional networks (convnets). Beyond the fact that adapting self-supervised methods to this architecture works particularly well, we make the following observations: first, self-supervised ViT features contain explicit information about the semantic segmentation of an image, which does not emerge as clearly with supervised ViTs, nor with convnets. Second, these features are also excellent k-NN classifiers, reaching 78.3% top-1 on ImageNet with a small ViT. Our study also underlines the importance of momentum encoder, multi-crop training, and the use of small patches with ViTs. We implement our findings into a simple self-supervised method, called DINO, which we interpret as a form of self-distillation with no labels. We show the synergy between DINO and ViTs by achieving 80.1% top-1 on ImageNet in linear evaluation with ViT-Base.

Mahmoud Assran, Mathilde Caron, Ishan Misra et al.

This paper proposes a novel method of learning by predicting view assignments with support samples (PAWS). The method trains a model to minimize a consistency loss, which ensures that different views of the same unlabeled instance are assigned similar pseudo-labels. The pseudo-labels are generated non-parametrically, by comparing the representations of the image views to those of a set of randomly sampled labeled images. The distance between the view representations and labeled representations is used to provide a weighting over class labels, which we interpret as a soft pseudo-label. By non-parametrically incorporating labeled samples in this way, PAWS extends the distance-metric loss used in self-supervised methods such as BYOL and SwAV to the semi-supervised setting. Despite the simplicity of the approach, PAWS outperforms other semi-supervised methods across architectures, setting a new state-of-the-art for a ResNet-50 on ImageNet trained with either 10% or 1% of the labels, reaching 75.5% and 66.5% top-1 respectively. PAWS requires 4x to 12x less training than the previous best methods.

Pedro Morgado, Ishan Misra, Nuno Vasconcelos

We present a self-supervised learning method to learn audio and video representations. Prior work uses the natural correspondence between audio and video to define a standard cross-modal instance discrimination task, where a model is trained to match representations from the two modalities. However, the standard approach introduces two sources of training noise. First, audio-visual correspondences often produce faulty positives since the audio and video signals can be uninformative of each other. To limit the detrimental impact of faulty positives, we optimize a weighted contrastive learning loss, which down-weighs their contribution to the overall loss. Second, since self-supervised contrastive learning relies on random sampling of negative instances, instances that are semantically similar to the base instance can be used as faulty negatives. To alleviate the impact of faulty negatives, we propose to optimize an instance discrimination loss with a soft target distribution that estimates relationships between instances. We validate our contributions through extensive experiments on action recognition tasks and show that they address the problems of audio-visual instance discrimination and improve transfer learning performance.

Mandela Patrick, Yuki M. Asano, Bernie Huang et al.

The quality of the image representations obtained from self-supervised learning depends strongly on the type of data augmentations used in the learning formulation. Recent papers have ported these methods from still images to videos and found that leveraging both audio and video signals yields strong gains; however, they did not find that spatial augmentations such as cropping, which are very important for still images, work as well for videos. In this paper, we improve these formulations in two ways unique to the spatio-temporal aspect of videos. First, for space, we show that spatial augmentations such as cropping do work well for videos too, but that previous implementations, due to the high processing and memory cost, could not do this at a scale sufficient for it to work well. To address this issue, we first introduce Feature Crop, a method to simulate such augmentations much more efficiently directly in feature space. Second, we show that as opposed to naive average pooling, the use of transformer-based attention improves performance significantly, and is well suited for processing feature crops. Combining both of our discoveries into a new method, Space-Time Crop & Attend (STiCA) we achieve state-of-the-art performance across multiple video-representation learning benchmarks. In particular, we achieve new state-of-the-art accuracies of 67.0% on HMDB-51 and 93.1% on UCF-101 when pre-training on Kinetics-400.

Jure Zbontar, Li Jing, Ishan Misra et al.

Self-supervised learning (SSL) is rapidly closing the gap with supervised methods on large computer vision benchmarks. A successful approach to SSL is to learn embeddings which are invariant to distortions of the input sample. However, a recurring issue with this approach is the existence of trivial constant solutions. Most current methods avoid such solutions by careful implementation details. We propose an objective function that naturally avoids collapse by measuring the cross-correlation matrix between the outputs of two identical networks fed with distorted versions of a sample, and making it as close to the identity matrix as possible. This causes the embedding vectors of distorted versions of a sample to be similar, while minimizing the redundancy between the components of these vectors. The method is called Barlow Twins, owing to neuroscientist H. Barlow's redundancy-reduction principle applied to a pair of identical networks. Barlow Twins does not require large batches nor asymmetry between the network twins such as a predictor network, gradient stopping, or a moving average on the weight updates. Intriguingly it benefits from very high-dimensional output vectors. Barlow Twins outperforms previous methods on ImageNet for semi-supervised classification in the low-data regime, and is on par with current state of the art for ImageNet classification with a linear classifier head, and for transfer tasks of classification and object detection.

Zaiwei Zhang, Rohit Girdhar, Armand Joulin et al.

Pretraining on large labeled datasets is a prerequisite to achieve good performance in many computer vision tasks like 2D object recognition, video classification etc. However, pretraining is not widely used for 3D recognition tasks where state-of-the-art methods train models from scratch. A primary reason is the lack of large annotated datasets because 3D data is both difficult to acquire and time consuming to label. We present a simple self-supervised pertaining method that can work with any 3D data - single or multiview, indoor or outdoor, acquired by varied sensors, without 3D registration. We pretrain standard point cloud and voxel based model architectures, and show that joint pretraining further improves performance. We evaluate our models on 9 benchmarks for object detection, semantic segmentation, and object classification, where they achieve state-of-the-art results and can outperform supervised pretraining. We set a new state-of-the-art for object detection on ScanNet (69.0% mAP) and SUNRGBD (63.5% mAP). Our pretrained models are label efficient and improve performance for classes with few examples.

Yutong Bai, Haoqi Fan, Ishan Misra et al.

Leveraging temporal information has been regarded as essential for developing video understanding models. However, how to properly incorporate temporal information into the recent successful instance discrimination based contrastive self-supervised learning (CSL) framework remains unclear. As an intuitive solution, we find that directly applying temporal augmentations does not help, or even impair video CSL in general. This counter-intuitive observation motivates us to re-design existing video CSL frameworks, for better integration of temporal knowledge. To this end, we present Temporal-aware Contrastive self-supervised learningTaCo, as a general paradigm to enhance video CSL. Specifically, TaCo selects a set of temporal transformations not only as strong data augmentation but also to constitute extra self-supervision for video understanding. By jointly contrasting instances with enriched temporal transformations and learning these transformations as self-supervised signals, TaCo can significantly enhance unsupervised video representation learning. For instance, TaCo demonstrates consistent improvement in downstream classification tasks over a list of backbones and CSL approaches. Our best model achieves 85.1% (UCF-101) and 51.6% (HMDB-51) top-1 accuracy, which is a 3% and 2.4% relative improvement over the previous state-of-the-art.