Abstract The pursuit of artificial general intelligence (AGI) has been accelerated by Multimodal Large Language Models (MLLMs), which exhibit superior reasoning, generalization capabilities, and proficiency in processing multimodal inputs. A crucial milestone in the evolution of AGI is the attainment of human-level planning, a fundamental ability for making informed decisions in complex environments, and solving a wide range of real-world problems. Despite the impressive advancements in MLLMs, a question remains: How far are current MLLMs from achieving human-level planning? To shed light on this question, we introduce EgoPlan-Bench, a comprehensive benchmark to evaluate the planning abilities of MLLMs in real-world scenarios from an egocentric perspective, mirroring human perception. EgoPlan-Bench emphasizes the evaluation of planning capabilities of MLLMs, featuring realistic tasks, diverse action plans, and intricate visual observations. Our rigorous evaluation of a wide range of MLLMs reveals that EgoPlan-Bench poses significant challenges, highlighting a substantial scope for improvement in MLLMs to achieve human-level task planning. To facilitate this advancement, we further present EgoPlan-IT, a specialized instruction-tuning dataset that effectively enhances model performance on EgoPlan-Bench. We have made all the codes, data, and a maintained benchmark leaderboard available at https://chenyi99.github.io/ego_plan/ to advance future research.

Abstract Multimodal Large Language Models (MLLMs) has recently demonstrated superior multimodal comprehension abilities, heralding a new era for artificial general intelligence (AGI). However, achieving AGI necessitates more than just comprehension. A crucial capability required is effective planning in diverse scenarios, which involves making reasonable decisions based on complex environments to solve real-world problems. Despite its importance, the planning abilities of current MLLMs in varied scenarios remain underexplored, leaving a significant gap in our understanding of their full potential. In this paper, we introduce EgoPlan-Bench2, a rigorous and comprehensive benchmark designed to assess the planning capabilities of MLLMs across a wide range of real-world scenarios . EgoPlan-Bench2 encompasses everyday tasks spanning 4 major domains and 24 detailed scenarios, closely aligned with human daily life. It is constructed through a semi-automatic process utilizing egocentric videos, complemented by manual verification. Grounded in a first-person perspective, it mirrors the way humans approach problem-solving in everyday life. We evaluate 25 competitive MLLMs and provide an in-depth analysis of their limitations, revealing that they face significant challenges in real-world planning. To diagnose the underlying bottlenecks, we investigate the effectiveness of various prompts via a training-free multimodal prompting method. We find that MLLMs’ planning performance on EgoPlan-Bench2 is critically dependent on temporally structured action sequences in historical task progress and interactions between objects and humans in current observation state. This dependency also underscores the necessity for strong reasoning abilities to integrate diverse multimodal cues and analysis before making final decision. Building on this insight, we demonstrate that EgoPlan-Bench2 is also an effective video reasoning benchmark. Experiments with Gemini-2.5-Flash and a post-trained Qwen-2.5-VL confirm its ability to distinguish between models with and without explicit deliberate reasoning mechanisms, showcasing the tangible impact of DeepSeek-R1 paradigm reasoning on planning tasks. We have made data and code available at https://qiulu66.github.io/egoplanbench2/ .

In recent years, there has been a significant surge of interest in unifying image comprehension and generation within Large Language Models (LLMs). This growing interest has prompted us to explore extending this unification to videos. The core challenge lies in developing a versatile video tokenizer that captures both the spatial characteristics and temporal dynamics of videos to obtain representations for LLMs, and the representations can be further decoded into realistic video clips to enable video generation. In this work, we introduce Divot, a Diffusion-Powered Video Tokenizer, which leverages the diffusion process for self-supervised video representation learning. We posit that if a video diffusion model can effectively de-noise video clips by taking the features of a video tokenizer as the condition, then the tokenizer has successfully captured robust spatial and temporal information. Additionally, the video diffusion model inherently functions as a de-tokenizer, decoding videos from their representations. Building upon the Divot tokenizer, we present Divot-LLM through video-to-text auto-regression and text-to-video generation by modeling the distributions of continuous-valued Divot features with a Gaussian Mixture Model. Experimental results demonstrate that our diffusion-based video tokenizer, when integrated with a pre-trained LLM, achieves competitive performance across various video comprehension and generation benchmarks. The instruction tuned Divot-LLM also excels in video storytelling, generating interleaved narratives and corresponding videos. Models and codes are available at https://github.com/TencentARC/Divot.

Large Vision Language Models (LVLMs) have achieved significant success across multi-modal tasks. However, the computational cost of processing long visual tokens can be prohibitively expensive on resource-limited devices. Previous methods have identified redundancy in visual tokens within the Large Language Model (LLM) decoder layers and have mitigated this by pruning tokens using a predefined or fixed ratio, thereby reducing computational over-head. Nonetheless, we observe that the impact of pruning ratio varies across different LLM layers and instances (image-prompt pairs). Therefore, it is essential to develop a layer-wise and instance-wise vision token pruning strategy to balance computational cost and model performance effectively. We propose ATP-LLaVA, a novel approach that adaptively determines instance-specific token pruning ratios for each LLM layer. Specifically, we introduce an Adaptive Token Pruning (ATP) module, which computes the importance score and pruning threshold based on input instance adaptively. The ATP module can be seamlessly integrated between any two LLM layers with negligible computational overhead. Additionally, we develop a Spatial Augmented Pruning (SAP) strategy that prunes visual tokens with both token redundancy and spatial modeling perspectives. Our approach reduces the average token count by 75% while maintaining performance, with only a minimal 1.9% degradation across seven widely used benchmarks.

Vision-Language Models (VLMs) have achieved remarkable success in various multi-modal tasks, but they are often bottlenecked by the limited context window and high computational cost of processing high-resolution image inputs and videos. Vision compression can alleviate this problem by reducing the vision token count. Previous approaches compress vision tokens with external modules and force LLMs to understand the compressed ones, leading to visual information loss. However, the LLMs’ understanding paradigm of vision tokens is not fully utilised in the compression learning process. We propose VoCo-LLaMA, the first approach to compress vision tokens using LLMs. By introducing Vision Compression tokens during the vision instruction tuning phase and leveraging attention distillation, our method distill how LLMs comprehend vision tokens into their processing of VoCo tokens. VoCo-LLaMA facilitates effective vision compression and improves the computational efficiency during the inference stage. Specifically, our method can achieve a 576× compression rate while maintaining 83.7% performance. Furthermore, through continuous training using time-series compressed token sequences of video frames, VoCo-LLaMA demonstrates the ability to understand temporal correlations, outperforming previous methods on popular video question-answering benchmarks. Our approach presents a promising way to unlock the full potential of VLMs’ contextual window, enabling more scalable multi-modal applications. The project page can be accessed via https://yxxxb.github.io/VoCo-LLaMA-page.

Recent advancements in image and video synthesis have opened up new promise in generative games. One particularly intriguing application is transforming characters from anime films into interactive, playable entities. This allows players to immerse themselves in the dynamic anime world as their favorite characters for life simulation through language instructions. Such games are defined as infinite game since they eliminate predetermined boundaries and fixed gameplay rules, where players can interact with the game world through open-ended language and experience ever-evolving storylines and environments. Recently, a pioneering approach for infinite anime life simulation employs large language models (LLMs) to translate multi-turn text dialogues into language instructions for image generation. However, it neglects historical visual context, leading to inconsistent gameplay. Furthermore, it only generates static images, failing to incorporate the dynamics necessary for an engaging gaming experience. In this work, we propose AnimeGamer, which is built upon Multimodal Large Language Models (MLLMs) to generate each game state, including dynamic animation shots that depict character movements and updates to character states, as illustrated in Figure 1. We introduce novel action-aware multimodal representations to represent animation shots, which can be decoded into high-quality video clips using a video diffusion model. By taking historical animation shot representations as context and predicting subsequent representations, AnimeGamer can generate games with contextual consistency and satisfactory dynamics. Extensive evaluations using both automated metrics and human evaluations demonstrate that AnimeGamer outperforms existing methods in various aspects of the gaming experience. Codes and checkpoints are available at https://github.com/TencentARC/AnimeGamer.

The synergy between generative and discriminative models receives growing attention. While discriminative Contrastive Language-Image Pre-Training (CLIP) excels in high-level semantics, it struggles with perceiving fine-grained visual details. Generally, to enhance representations, generative models take CLIP's visual features as conditions for reconstruction. However, the underlying principle remains underexplored. In this work, we empirically found that visually perfect generations are not always optimal for representation enhancement. The essence lies in effectively extracting fine-grained knowledge from generative models while mitigating irrelevant information. To explore critical factors, we delve into three aspects: (1) Conditioning mechanisms: We found that even a small number of local tokens can drastically reduce the difficulty of reconstruction, leading to collapsed training. We thus conclude that utilizing only global visual tokens as conditions is the most effective strategy. (2) Denoising configurations: We observed that end-to-end training introduces extraneous information. To address this, we propose a two-stage training strategy to prioritize learning useful visual knowledge. Additionally, we demonstrate that lightweight denoisers can yield remarkable improvements. (3) Generation paradigms: We explore both continuous and discrete denoisers with desirable outcomes, validating the versatility of our method. Through our in-depth explorations, we have finally arrived at an effective method, namely GenHancer, which consistently outperforms prior arts on the MMVP-VLM benchmark, e.g., 6.0% on OpenAICLIP. The enhanced CLIP can be further plugged into multimodal large language models for better vision-centric performance. All the models and codes are made publicly available.

Recent advancements in large language models (LLMs) have significantly propelled the development of large multi-modal models (LMMs), highlighting the potential for general and intelligent assistants. However, most LMMs model visual and textual modalities separately, leading to recent efforts to develop native LMMs using a single transformer. Despite the promise, these native models are resource-intensive and often exhibit performance gaps compared to their compositional counterparts. To alleviate this issue, we propose a simple yet efficient method to construct a baseline for the native and end-to-end large multi-modal model in a single transformer. First, we propose a new early-fusion LMM that can fuse multi-modal inputs in the early stage and respond to visual instructions in an auto-regressive manner. Second, we devise an efficient training recipe for the proposed model, which harnesses the prior knowledge of the pre-trained models, addressing both the performance limitations and the challenge of resource consumption. The proposed model demonstrates superior performance compared to other LMMs using one transformer and significantly narrows the performance gap with compositional LMMs.

Recent advances have highlighted the benefits of scaling language models to enhance performance across a wide range of NLP tasks. However, these approaches still face limitations in effectiveness and efficiency when applied to domain-specific tasks, particularly for small edge-side models. We propose the LoRA-Gen framework, which utilizes a large cloud-side model to generate LoRA parameters for edge-side models based on task descriptions. By employing the reparameterization technique, we merge the LoRA parameters into the edge-side model to achieve flexible specialization. Our method facilitates knowledge transfer between models while significantly improving the inference efficiency of the specialized model by reducing the input context length. Without specialized training, LoRA-Gen outperforms conventional LoRA fine-tuning, which achieves competitive accuracy and a 2.1x speedup with TinyLLaMA-1.1B in reasoning tasks. Besides, our method delivers a compress ratio of 10.1x with Gemma-2B on intelligent agent tasks.

Range-only Simultaneous Localisation and Mapping (RO-SLAM) is of interest due to its practical applications in ultra-wideband (UWB) and Bluetooth Low Energy (BLE) localisation in terrestrial and aerial applications and acoustic beacon localisation in submarine applications. In this work, we consider a mobile robot equipped with an inertial measurement unit (IMU) and a range sensor that measures distances to a collection of fixed landmarks. We derive an equivariant filter (EqF) for the RO-SLAM problem based on a symmetry Lie group that is compatible with the range measurements. The proposed filter does not require bootstrapping or initialisation of landmark positions, and demonstrates robustness to the noprior situation. The filter is demonstrated on a real-world dataset, and it is shown to significantly outperform a state-of-the-art EKF alternative in terms of both accuracy and robustness.

This work introduces a new Transformer model called Cached Transformer, which uses Gated Recurrent Cached (GRC) attention to extend the self-attention mechanism with a differentiable memory cache of tokens. GRC attention enables attending to both past and current tokens, increasing the receptive field of attention and allowing for exploring long-range dependencies. By utilizing a recurrent gating unit to continuously update the cache, our model achieves significant advancements in \textbf{six} language and vision tasks, including language modeling, machine translation, ListOPs, image classification, object detection, and instance segmentation. Furthermore, our approach surpasses previous memory-based techniques in tasks such as language modeling and displays the ability to be applied to a broader range of situations.

Chengyue Wu, Yukang Gan, Yixiao Ge, Zeyu Lu, Jiahao Wang, Ye Feng, Ying Shan, Ping Luo. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2024.

The You Only Look Once (YOLO) series of detectors have established themselves as efficient and practical tools. However, their reliance on predefined and trained object categories limits their applicability in open scenarios. Addressing this limitation, we introduce YOLO-World, an innovative approach that enhances YOLO with open-vocabulary detection capabilities through vision-language modeling and pre-training on large-scale datasets. Specifically, we propose a new Re-parameterizable Vision-Language Path Aggregation Network (RepVL-PAN) and region-text contrastive loss to facilitate the interaction between visual and linguistic information. Our method excels in detecting a wide range of objects in a zero-shot manner with high efficiency. On the challenging LVIS dataset, YOLO-World achieves 35.4 AP with 52.0 FPS on V100, which outperforms many state-of-the-art methods in terms of both accuracy and speed. Furthermore, the finetuned YOLO-World achieves remarkable performance on several downstream tasks, including object detection and open-vocabulary instance segmentation. Code and models are available at: https://github.com/AILab-eve/YOLO-World.

Large-kernel convolutional neural networks (ConvNets) have recently received extensive research attention, but two unresolved and critical issues demand further investigation. 1) The architectures of existing large-kernel ConvNets largely follow the design principles of conventional ConvNets or transformers, while the architectural design for large-kernel ConvNets remains under-addressed. 2) As transformers have dominated multiple modalities, it re-mains to be investigated whether ConvNets also have a strong universal perception ability in domains beyond vision. In this paper, we contribute from two aspects. 1) We propose four architectural guidelines for designing large- kernel ConvNets, the core of which is to exploit the essential characteristics of large kernels that distinguish them from small kernels - they can see wide without going deep. Fol-lowing such guidelines, our proposed large-kernel ConvNet shows leading performance in image recognition (ImageNet accuracy of 88.0%, ADE20K mIoU of 55.6%, and COCO box AP of 56.4%), demonstrating better performance and higher speed than the recent powerful competitors. 2) We discover large kernels are the key to unlocking the exceptional performance of ConvNets in domains where they were originally not proficient. With certain modality-related pre-processing approaches, the proposed model achieves state- of-the-art performance on time-series forecasting and audio recognition tasks even without modality-specific customization to the architecture. All the code and models are publicly available on GitHub and Huggingface.

Vector-Quantized (VQ-based) generative models usually consist of two basic components, i.e., VQ tokenizers and generative transformers. Prior research focuses on improving the reconstruction fidelity of VQ tokenizers but rarely examines how the improvement in reconstruction affects the generation ability of generative transformers. In this paper, we surprisingly find that improving the reconstruction fidelity of VQ tokenizers does not necessarily improve the generation. Instead, learning to compress semantic features within VQ tokenizers significantly improves generative transformers' ability to capture textures and structures. We thus highlight two competing objectives of VQ tokeniz-ers for image synthesis: semantic compression and details preservation. Different from previous work that pri-oritizes better details preservation, we propose Semantic-Quantized GAN (SeQ-GAN) with two learning phases to balance the two objectives. In the first phase, we propose a semantic-enhanced perceptual loss for better semantic compression. In the second phase, we fix the encoder and codebook, but enhance and finetune the decoder to achieve better details preservation. Our proposed SeQ-GAN significantly improves VQ-based generative models for both un-conditional and conditional image generation. Specifically, SeQ-GAN achieves a Fré chet Inception Distance (FID) of 6.25 and Inception Score (IS) of 140.9 on 256×256 Ima-geNet generation, which is a remarkable improvement over VIT-VQGAN (714M), which obtains 11.2 FID and 97.2 IS.

Current instruction-based image editing methods, such as InstructPix2Pix, often fail to produce satisfactory results in complex scenarios due to their dependence on the simple CLIP text encoder in diffusion models. To rectify this, this paper introduces SmartEdit, a novel approach of instruction-based image editing that leverages Multimodal Large Language Models (MLLMs) to enhance its understanding and reasoning capabilities. However, direct integration of these elements still faces challenges in situations requiring complex reasoning. To mitigate this, we propose a Bidirectional Interaction Module (BIM) that enables comprehensive bidirectional information interactions between the input image and the MLLM output. During training, we initially incorporate perception data to boost the perception and understanding capabilities of diffusion models. Subsequently, we demonstrate that a small amount of complex instruction editing data can effectively stimulate SmartEdit’ s editing capabilities for more complex instructions. We further construct a new evaluation dataset, Reason-Edit, specifically tailored for complex instruction-based image editing. Both quantitative and qualitative results on this evaluation dataset indicate that our SmartEdit surpasses previous methods, paving the way for the practical application of complex instruction-based image editing.

Aiming to advance AI agents, large foundation models significantly improve reasoning and instruction execution, yet the current focus on vision and language neglects the potential of perceiving diverse modalities in open-world environments. However, the success of data-driven vision and language models is costly or even infeasible to be reproduced for rare modalities. In this paper, we present Vit-lens that facilitates efficient omni-modal representation learning by perceiving novel modalities with a pretrained- ViT and aligning them to a pre-defined space. Specifically, the modality-specific lens is tuned to project any-modal signals to an intermediate embedding space, which are then processed by a strong ViT with pre-trained visual knowledge. The encoded representations are optimized toward aligning with the modal-independent space, pre-defined by off-the-shelf foundation models. Vit-lensprovides a unified solution for representation learning of increasing modalities with two appealing advantages: (i) Unlocking the great potential of pretrained- ViTs to novel modalities effectively with efficient parameters and data regime; (ii) Enabling emergent down- stream capabilities through modality alignment and shared ViT parameters. We tailor Vit-lensto learn representations for 3D point cloud, depth, audio, tactile and EEG, and set new state-of-the-art results across various understanding tasks, such as zero-shot classification. By seamlessly integrating Vit-lensinto Multimodal Foundation Models, we enable Any-modality to Text and Image Generation in a zero-shot manner. Code and models are available at https://github.com/TencentARC/ViT-Lens.

Multimodal large language models (MLLMs), building upon the foundation of powerful large language models (LLMs), have recently demonstrated exceptional capabilities in generating not only texts but also images given in-terleaved multimodal inputs (acting like a combination of GPT-4V and DALL-E 3). However, existing MLLM benchmarks remain limited to assessing only models' comprehension ability of single image-text inputs, failing to keep up with the strides made in MLLMs. A comprehensive benchmark is imperative for investigating the progress and uncovering the limitations of current MLLMs. In this work, we categorize the capabilities of MLLMs into hierarchical levels from L0to L4based on the modalities they can ac-cept and generate, and propose SEED-Bench, a comprehensive benchmark that evaluates the hierarchical capa-bilities of MLLMs. Specifically, SEED-Bench comprises 24K multiple-choice questions with accurate human annotations, which span 27 dimensions, including the evaluation of both text and image generation. Multiple-choice questions with ground truth options derived from human annotation enable an objective and efficient assessment of model performance, eliminating the need for human or GPT intervention during evaluation. We further evaluate the performance of 22 prominent open-source MLLMs and summarize valuable observations. By revealing the limitations of existing MLLMs through extensive evaluations, we aim for SEED-Bench to provide insights that will mo-tivate future research toward the goal of General Artificial Intelligence. Dataset and evaluation code are available at https://github.com/AILab-CVC/SEED-Bench.

The recent progress in Large Language Models (LLM) has spurred various advancements in image-language con-versation agents, while how to build a proficient video-based dialogue system is still under exploration. Consid-ering the extensive scale of LLM and visual backbone, min-imal GPU memory is left for facilitating effective temporal modeling, which is crucial for comprehending and providing feedback on videos. To this end, we propose Branching Temporal Adapter (BT-Adapter), a novel method for ex-tending image-language pretrained models into the video domain. Specifically, BT-Adapter serves as a plug-and-use temporal modeling branch alongside the pretrained vi-sual encoder, which is tuned while keeping the backbone frozen. Just pretrained once, BT-Adapter can be seamlessly integrated into all image conversation models using this version of CLIP, enabling video conversations without the need for video instructions. Besides, we develop a unique asymmetric token masking strategy inside the branch with tailor-made training tasks for BT-Adapter, facilitating faster convergence and better results. Thanks to BT-Adapter, we are able to empower existing multimodal dialogue models with strong video understanding capabilities without incur-ring excessive GPU costs. Without bells and whistles, BT-Adapter achieves (1) state-of-the-art zero-shot results on various video tasks using thousands of fewer GPU hours. (2) better performance than current video chatbots without any video instruction tuning. (3) state-of-the-art results of video chatting using video instruction tuning, outperforming previous SOTAs by a large margin. The code has been available at https://github.com/farewellthreeIBT-Adapter.

This paper focuses on the high computational complexity in Large Language Models (LLMs), a significant challenge in both natural language processing (NLP) and multi-modal tasks. We propose Low-Rank Approximation for Sparse Attention (LoRA -Sparse), an innovative approach that strategically reduces this complexity. LoRA -Sparse introduces low-rank linear projection layers for sparse attention approximation. It utilizes an order-mimic training methodology, which is crucial for efficiently approximating the self-attention mechanism in LLMs. We empirically show that sparse attention not only reduces computational demands, but also enhances model performance in both NLP and multi-modal tasks. This surprisingly shows that redundant attention in LLMs might be non-beneficial. We extensively validate LoRA -Sparse through rigorous empirical studies in both (NLP) and multi-modal tasks, demonstrating its effectiveness and general applicability. Based on LLaMA and LLaVA models, our methods can reduce more than half of the self-attention computation with even better performance than full-attention baselines.

We propose to improve transformers of a specific modality with irrelevant data from other modalities, e.g., improve an ImageNet model with audio or point cloud datasets. We would like to highlight that the data samples of the target modality are irrelevant to the other modalities, which distinguishes our method from other works utilizing paired (e.g., CLIP) or interleaved data of different modalities. We propose a methodology named Multimodal Pathway - given a target modality and a transformer designed for it, we use an auxiliary transformer trained with data of another modality and construct pathways to connect components of the two models so that data of the target modality can be processed by both models. In this way, we utilize the universal sequence-to-sequence modeling abilities of transformers obtained from two modalities. As a concrete implementation, we use a modality-specific tokenizer and task-specific head as usual but utilize the transformer blocks of the auxiliary model via a proposed method named Cross-Modal Re-parameterization, which exploits the auxiliary weights without any inference costs. On the image, point cloud, video, and audio recognition tasks, we observe significant and consistent performance improvements with irrelevant data from other modalities. The code and models are available at https://github.com/AILab-CVC/M2PT.

The great success of Large Language Models (LLMs) has expanded the potential of multimodality, contributing to the gradual evolution of General Artificial Intelligence (AGI). A true AGI agent should not only possess the capability to perform predefined multi-tasks but also exhibit emergent abilities in an open-world context. However, despite the considerable advancements made by recent multimodal LLMs, they still fall short in effectively unifying comprehension and generation tasks, let alone open-world emergent abilities. We contend that the key to overcoming the present impasse lies in enabling text and images to be represented and processed interchangeably within a unified autoregressive Transformer. To this end, we introduce $\textbf{SEED}$, an elaborate image tokenizer that empowers LLMs with the ability to $\textbf{SEE}$ and $\textbf{D}$raw at the same time. We identify two crucial design principles: (1) Image tokens should be independent of 2D physical patch positions and instead be produced with a $\textit{1D causal dependency}$, exhibiting intrinsic interdependence that aligns with the left-to-right autoregressive prediction mechanism in LLMs. (2) Image tokens should capture $\textit{high-level semantics}$ consistent with the degree of semantic abstraction in words, and be optimized for both discriminativeness and reconstruction during the tokenizer training phase. With SEED tokens, LLM is able to perform scalable multimodal autoregression under its original training recipe, i.e., next-word prediction. SEED-LLaMA is therefore produced by large-scale pretraining and instruction tuning on the interleaved textual and visual data, demonstrating impressive performance on a broad range of multimodal comprehension and generation tasks. More importantly, SEED-LLaMA has exhibited compositional emergent abilities such as multi-turn in-context multimodal generation, acting like your AI assistant. The code (training and inference) and models are released in https://github.com/AILab-CVC/SEED.

The majority of commercial and open-source autopilot software for uncrewed aerial vehicles rely on the tried and tested extended Kalman filter (EKF) to provide the state estimation solution for the inertial navigation system (INS). While modern implementations achieve remarkable robustness, it is often due to the careful implementation of exception code for a multitude of corner cases along with significant skilled tuning effort. In this paper, we use the data wealth of the ArduPilot community to identify and highlight the most common real-world challenges in INS state estimation, including sensor self-calibration, robustness in static conditions, global navigation satellite system (GNSS) outliers and shifts, and robustness to faulty inertial measurement units (IMUs). We propose a novel equivariant filter (EqF) formulation for the INS solution that exploits a Semi-Direct-Bias symmetry group for multi-sensor fusion with self-calibration capabilities and incorporates equivariant velocity-type measurements. We augment the filter with a simple innovation-covariance inflation strategy that seamlessly handles GNSS outliers and shifts without requiring coding of a whole set of exception cases. We use real-world data from the Ardupilot community to demonstrate the performance of the proposed filter on known cases where existing filters fail without careful exception handling or case-specific tuning and benchmark against the ArduPilot’s EKF3, the most sophisticated EKF implementation currently available.

The state space models, employing recursively propagated features, demonstrate strong representation capabilities comparable to Transformer models and superior efficiency. However, constrained by the inherent geometric constraints of sequences, it still falls short in modeling long-range dependencies. To address this issue, we propose the MambaTree network, which first dynamically generates a tree topology based on spatial relationships and input features. Then, feature propagation is performed based on this graph, thereby breaking the original sequence constraints to achieve stronger representation capabilities. Additionally, we introduce a linear complexity dynamic programming algorithm to enhance long-range interactions without increasing computational cost. MambaTree is a versatile multimodal framework that can be applied to both visual and textual tasks. Extensive experiments demonstrate that our method significantly outperforms existing structured state space models on image classification, object detection and segmentation. Besides, by fine-tuning large language models, our approach achieves consistent improvements in multiple textual tasks at minor training cost.

Unsupervised domain adaptation (UDA) aims at adapting the model trained on a labeled source-domain dataset to an unlabeled target-domain dataset. The task of UDA on open-set person reidentification (re-ID) is even more challenging as the identities (classes) do not have overlap between the two domains. One major research direction was based on domain translation, which, however, has fallen out of favor in recent years due to inferior performance compared with pseudo-label-based methods. We argue that domain translation has great potential on exploiting valuable source-domain data but the existing methods did not provide proper regularization on the translation process. Specifically, previous methods only focus on maintaining the identities of the translated images while ignoring the intersample relations during translation. To tackle the challenges, we propose an end-to-end structured domain adaptation framework with an online relation-consistency regularization term. During training, the person feature encoder is optimized to model intersample relations on-the-fly for supervising relation-consistency domain translation, which in turn improves the encoder with informative translated images. The encoder can be further improved with pseudo labels, where the source-to-target translated images with ground-truth identities and target-domain images with pseudo identities are jointly used for training. In the experiments, our proposed framework is shown to achieve state-of-the-art performance on multiple UDA tasks of person re-ID. With the synthetic→real translated images from our structured domain-translation network, we achieved second place in the Visual Domain Adaptation Challenge (VisDA) in 2020.

Video-Text pre-training aims at learning transferable representations from large-scale video-text pairs via aligning the semantics between visual and textual information. State-of-the-art approaches extract visual features from raw pixels in an end-to-end fashion. However, these methods operate at frame-level directly and thus overlook the spatio-temporal structure of objects in video, which yet has a strong synergy with nouns in textual descriptions. In this work, we propose a simple yet effective module for video-text representation learning, namely RegionLearner, which can take into account the structure of objects during pre-training on large-scale video-text pairs. Given a video, our module (1) first quantizes continuous visual features via clustering patch-features into the same cluster according to content similarity, then (2) generates learnable masks to aggregate fragmentary features into regions with complete semantics, and finally (3) models the spatio-temporal dependencies between different semantic regions. In contrast to using off-the-shelf object detectors, our proposed module does not require explicit supervision and is much more computationally efficient. We pre-train the proposed approach on the public WebVid2M and CC3M datasets. Extensive evaluations on four downstream video-text retrieval benchmarks clearly demonstrate the effectiveness of our RegionLearner.

The traditional model upgrading paradigm for retrieval requires recomputing all gallery embeddings before deploying the new model (dubbed as "backfilling"), which is quite expensive and time-consuming considering billions of instances in industrial applications. BCT presents the first step towards backward-compatible model upgrades to get rid of backfilling. It is workable but leaves the new model in a dilemma between new feature discriminativeness and new-to-old compatibility due to the undifferentiated compatibility constraints. In this work, we propose Darwinian Model Upgrades (DMU), which disentangle the inheritance and variation in the model evolving with selective backward compatibility and forward adaptation, respectively. The old-to-new heritable knowledge is measured by old feature discriminativeness, and the gallery features, especially those of poor quality, are evolved in a lightweight manner to become more adaptive in the new latent space. We demonstrate the superiority of DMU through comprehensive experiments on large-scale landmark retrieval and face recognition benchmarks. DMU effectively alleviates the new-to-new degradation at the same time improving new-to-old compatibility, rendering a more proper model upgrading paradigm in large-scale retrieval systems.Code: https://github.com/TencentARC/OpenCompatible.

Mainstream Video-Language Pre-training (VLP) models [10, 26, 64] consist of three parts, a video encoder, a text encoder, and a video-text fusion Transformer. They pursue better performance via utilizing heavier unimodal encoders or multimodal fusion Transformers, resulting in increased parameters with lower efficiency in downstream tasks. In this work, we for the first time introduce an end-to-end VLP model, namely all-in-one Transformer, that embeds raw video and textual signals into joint representations using a unified backbone architecture. We argue that the unique temporal information of video data turns out to be a key barrier hindering the design of a modality-agnostic Transformer. To overcome the challenge, we introduce a novel and effective token rolling operation to encode temporal representations from video clips in a non-parametric manner. The careful design enables the representation learning of both video-text multimodal inputs and unimodal inputs using a unified model. Our pretrained ali-in-one Transformer is transferred to various downstream video-text tasks after fine-tuning, including text-video retrieval, video-question answering, multiple choice and video captioning. State-of-the-art performances with the minimal model FLOPs on ten datasets demonstrate the superiority of our method compared to the competitive counterparts. The code and pretrained models are available at https://github.com/showlab/all-in-one.

The state of the arts in vision-language pretraining (VLP) achieves exemplary performance but suffers from high training costs resulting from slow convergence and long training time, especially on large-scale web datasets. An essential obstacle to training efficiency lies in the entangled prediction rate (percentage of tokens for reconstruction) and corruption rate (percentage of corrupted tokens) in masked language modeling (MLM), that is, a proper corruption rate is achieved at the cost of a large portion of output tokens being excluded from prediction loss. To accelerate the convergence of VLP, we propose a new pretraining task, namely, free language modeling (FLM), that enables a 100% prediction rate with arbitrary corruption rates. FLM successfully frees the prediction rate from the tie-up with the corruption rate while allowing the corruption spans to be customized for each token to be predicted. FLM-trained models are encouraged to learn better and faster given the same GPU time by exploiting bidirectional contexts more flexibly. Extensive experiments show FLM could achieve an impressive 2.5 × pretraining time reduction in comparison to the MLM-based methods, while keeping competitive performance on both vision-language understanding and generation tasks. Code will be public at https://github.com/TencentARC/FLM.

Both masked image modeling (MIM) and natural language supervision have facilitated the progress of transferable visual pre-training. In this work, we seek the synergy between two paradigms and study the emerging properties when MIM meets natural language supervision. To this end, we present a novel masked visual Reconstruction In Language semantic Space (RILS) pre-training framework, in which sentence representations, encoded by the text encoder, serve as prototypes to transform the vision-only signals into patch-sentence probabilities as semantically meaningful MIM reconstruction targets. The vision models can therefore capture useful components with structured information by predicting proper semantic of masked tokens. Better visual representations could, in turn, improve the text encoder via the image-text alignment objective, which is essential for the effective MIM target transformation. Extensive experimental results demonstrate that our method not only enjoys the best of previous MIM and CLIP but also achieves further improvements on various tasks due to their mutual benefits. RILS exhibits advanced transferability on downstream classification, detection, and segmentation, especially for low-shot regimes. Code is available at https://github.com/hustvl/RILS.

Pre-training on large-scale video data has become a common recipe for learning transferable spatiotemporal representations in recent years. Despite some progress, existing methods are mostly limited to highly curated datasets (e.g., K400) and exhibit unsatisfactory out-of-the-box representations. We argue that it is due to the fact that they only capture pixel-level knowledge rather than spatiotemporal semantics, which hinders further progress in video understanding. Inspired by the great success of image-text pre-training (e.g., CLIP), we take the first step to exploit language semantics to boost transferable spatiotemporal representation learning. We introduce a new pre-text task, Turning to Video for Transcript Sorting (TVTS), which sorts shuffled ASR scripts by attending to learned video representations. We do not rely on descriptive captions and learn purely from video, i.e., leveraging the natural transcribed speech knowledge to provide noisy but useful semantics over time. Our method enforces the vision model to contextualize what is happening over time so that it can re-organize the narrative transcripts, and can seamlessly apply to large-scale uncurated video data in the real world. Our method demonstrates strong out-of-the-box spatiotemporal representations on diverse benchmarks, e.g., +13.6% gains over VideoMAE on SSV2 via linear probing. The code is available at https://github.com/TencentARC/TVTS.

We present an approach to efficiently and effectively adapt a masked image modeling (MIM) pre-trained vanilla Vision Transformer (ViT) for object detection, which is based on our two novel observations: (i) A MIM pre-trained vanilla ViT encoder can work surprisingly well in the challenging object-level recognition scenario even with randomly sampled partial observations, e.g., only 25% ~ 50% of the input embeddings. (ii) In order to construct multi-scale representations for object detection from single-scale ViT, a randomly initialized compact convolutional stem supplants the pre-trained patchify stem, and its intermediate features can naturally serve as the higher resolution inputs of a feature pyramid network without further upsampling or other manipulations. While the pre-trained ViT is only regarded as the 3rd-stage of our detector’s backbone instead of the whole feature extractor. This naturally results in a ConvNet-ViT hybrid architecture. The proposed detector, named MimDet, enables a MIM pre-trained vanilla ViT to outperform leading hierarchical architectures such as Swin Transformer, MViTv2 and ConvNeXt on COCO object detection & instance segmentation, and achieves better results compared with the previous best adapted vanilla ViT detector using a more modest fine-tuning recipe while converging 2.8× faster. Code and pre-trained models are available at https://github.com/hustvl/MIMDet.

Transfer learning has become crucial in computer vision tasks due to the vast availability of pre-trained deep learning models. However, selecting the optimal pre-trained model from a diverse pool for a specific downstream task remains a challenge. Existing methods for measuring the transferability of pre-trained models rely on statistical correlations between encoded static features and task labels, but they overlook the impact of underlying representation dynamics during fine-tuning, leading to unreliable results, especially for self-supervised models. In this paper, we present an insightful physics-inspired approach named PED to address these challenges. We reframe the challenge of model selection through the lens of potential energy and directly model the interaction forces that influence fine-tuning dynamics. By capturing the motion of dynamic representations to decline the potential energy within a force-driven physical model, we can acquire an enhanced and more stable observation for estimating transferability. The experimental results on 10 downstream tasks and 12 self-supervised models demonstrate that our approach can seamlessly integrate into existing ranking techniques and enhance their performances, revealing its effectiveness for the model selection task and its potential for understanding the mechanism in transfer learning. Code is available at https://github.com/lixiaotong97/PED.

To replicate the success of text-to-image (T2I) generation, recent works employ large-scale video datasets to train a text-to-video (T2V) generator. Despite their promising results, such paradigm is computationally expensive. In this work, we propose a new T2V generation setting—One-Shot Video Tuning, where only one text-video pair is presented. Our model is built on state-of-the-art T2I diffusion models pre-trained on massive image data. We make two key observations: 1) T2I models can generate still images that represent verb terms; 2) extending T2I models to generate multiple images concurrently exhibits surprisingly good content consistency. To further learn continuous motion, we introduce Tune-A-Video, which involves a tailored spatio-temporal attention mechanism and an efficient one-shot tuning strategy. At inference, we employ DDIM inversion to provide structure guidance for sampling. Extensive qualitative and numerical experiments demonstrate the remarkable ability of our method across various applications.

Box-supervised instance segmentation has gained much attention as it requires only simple box annotations instead of costly mask or polygon annotations. However, existing box-supervised instance segmentation models mainly focus on mask-based frameworks. We propose a new end-to-end training technique, termed BoxSnake, to achieve effective polygonal instance segmentation using only box annotations for the first time. Our method consists of two loss functions: (1) a point-based unary loss that constrains the bounding box of predicted polygons to achieve coarse-grained segmentation; and (2) a distance-aware pairwise loss that encourages the predicted polygons to fit the object boundaries. Compared with the mask-based weakly-supervised methods, BoxSnake further reduces the performance gap between the predicted segmentation and the bounding box, and shows significant superiority on the Cityscapes dataset. The source code has been available at https://github.com/Yangr116/BoxSnake.

Large-scale embedding-based retrieval (EBR) is the cornerstone of search-related industrial applications. Given a user query, the system of EBR aims to identify relevant information from a large corpus of documents that may be tens or hundreds of billions in size. The storage and computation turn out to be expensive and inefficient with massive documents and high concurrent queries, making it difficult to further scale up.

The contrastive vision-language pre-training, known as CLIP, demonstrates remarkable potential in perceiving open-world visual concepts, enabling effective zero-shot image recognition. Nevertheless, few-shot learning methods based on CLIP typically require offline fine-tuning of the parameters on few-shot samples, resulting in longer inference time and the risk of overfitting in certain domains. To tackle these challenges, we propose the Meta-Adapter, a lightweight residual-style adapter, to refine the CLIP features guided by the few-shot samples in an online manner. With a few training samples, our method can enable effective few-shot learning capabilities and generalize to unseen data or tasks without additional fine-tuning, achieving competitive performance and high efficiency. Without bells and whistles, our approach outperforms the state-of-the-art online few-shot learning method by an average of 3.6\% on eight image classification datasets with higher inference speed. Furthermore, our model is simple and flexible, serving as a plug-and-play module directly applicable to downstream tasks. Without further fine-tuning, Meta-Adapter obtains notable performance improvements in open-vocabulary object detection and segmentation tasks.

Public large-scale text-to-image diffusion models, such as Stable Diffusion, have gained significant attention from the community. These models can be easily customized for new concepts using low-rank adaptations (LoRAs). However, the utilization of multiple-concept LoRAs to jointly support multiple customized concepts presents a challenge. We refer to this scenario as decentralized multi-concept customization, which involves single-client concept tuning and center-node concept fusion. In this paper, we propose a new framework called Mix-of-Show that addresses the challenges of decentralized multi-concept customization, including concept conflicts resulting from existing single-client LoRA tuning and identity loss during model fusion. Mix-of-Show adopts an embedding-decomposed LoRA (ED-LoRA) for single-client tuning and gradient fusion for the center node to preserve the in-domain essence of single concepts and support theoretically limitless concept fusion. Additionally, we introduce regionally controllable sampling, which extends spatially controllable sampling (e.g., ControlNet and T2I-Adapter) to address attribute binding and missing object problems in multi-concept sampling. Extensive experiments demonstrate that Mix-of-Show is capable of composing multiple customized concepts with high fidelity, including characters, objects, and scenes.

This paper aims to efficiently enable Large Language Models (LLMs) to use multi-modal tools. The advanced proprietary LLMs, such as ChatGPT and GPT-4, have shown great potential for tool usage through sophisticated prompt engineering. Nevertheless, these models typically rely on prohibitive computational costs and publicly inaccessible data. To address these challenges, we propose the GPT4Tools based on self-instruct to enable open-source LLMs, such as LLaMA and OPT, to use tools. It generates an instruction-following dataset by prompting an advanced teacher with various multi-modal contexts. By using the Low-Rank Adaptation (LoRA) optimization, our approach facilitates the open-source LLMs to solve a range of visual problems, including visual comprehension and image generation. Moreover, we provide a benchmark to evaluate the ability of LLMs to use tools, which is performed in both zero-shot and fine-tuning ways. Extensive experiments demonstrate the effectiveness of our method on various language models, which not only significantly improves the accuracy of invoking seen tools, but also enables the zero-shot capacity for unseen tools.

Pretraining a model to learn transferable video-text representation for retrieval has attracted a lot of attention in recent years. Previous dominant works mainly adopt two separate encoders for efficient retrieval, but ignore local associations between videos and texts. Another line of research uses a joint encoder to interact video with texts, but results in low efficiency since each text-video pair needs to be fed into the model. In this work, we enable fine-grained video-text interactions while maintaining high efficiency for retrieval via a novel pretext task, dubbed as Multiple Choice Questions (MCQ), where a parametric module BridgeFormer is trained to answer the “questions” constructed by the text features via resorting to the video features. Specifically, we exploit the rich semantics of text (i.e., nouns and verbs) to build questions, with which the video encoder can be trained to capture more regional content and temporal dynamics. In the form of questions and answers, the semantic associations between local video-text features can be properly established. BridgeFormer is able to be removed for downstream retrieval, rendering an efficient and flexible model with only two encoders. Our method outperforms state-of-the-art methods on the popular text-to-video retrieval task in five datasets with different experimental setups (i.e., zero-shot andfine-tune), including HowTo100M (one million videos). We further conduct zero-shot action recognition, which can be cast as video-to-text retrieval, and our approach also significantly surpasses its counterparts. As an additional benefit, our method achieves competitive results with much shorter pre-training videos on single-modality downstream tasks, e.g., action recognition with linear evaluation.

Recently, by introducing large-scale dataset and strong transformer network, video-language pre-training has shown great success especially for retrieval. Yet, existing video-language transformer models do not explicitly fine-grained semantic align. In this work, we present Object-aware Transformers, an object-centric approach that extends video-language transformer to incorporate object representations. The key idea is to leverage the bounding boxes and object tags to guide the training process. We evaluate our model on three standard sub-tasks of video-text matching on four widely used benchmarks. We also provide deep analysis and detailed ablation about the proposed method. We show clear improvement in performance across all tasks and datasets considered, demonstrating the value of a model that incorporates object representations into a video-language architecture. The code has been released in https://github.com/FingerRec/OA-Transformer.

Conventional model upgrades for visual search systems require offline refresh of gallery features by feeding gallery images into new models (dubbed as “backfill”), which is time-consuming and expensive, especially in large-scale applications. The task of backward-compatible representation learning is therefore introduced to support backfill-free model upgrades, where the new query features are interoperable with the old gallery features. Despite the success, previous works only investigated a close-set training scenario (i.e., the new training set shares the same classes as the old one), and are limited by more realistic and challenging open-set scenarios. To this end, we first introduce a new problem of universal backward-compatible representation learning, covering all possible data split in model upgrades. We further propose a simple yet effective method, dubbed as Universal Backward-Compatible Training (UniBCT) with a novel structural prototype refinement algorithm, to learn compatible representations in all kinds of model upgrading benchmarks in a unified manner. Comprehensive experiments on the large-scale face recognition datasets MS1Mv3 and IJB-C fully demonstrate the effectiveness of our method. Source code is available at https://github.com/TencentARC/OpenCompatible.

Conditional generative adversarial networks (cGANs) target at synthesizing diverse images given the input conditions and latent codes, but unfortunately, they usually suffer from the issue of mode collapse. To solve this issue, previous works [47], [22] mainly focused on encouraging the correlation between the latent codes and their generated images, while ignoring the relations between images generated from various latent codes. The recent MSGAN [27] tried to encourage the diversity of the generated image but only considers "negative" relations between the image pairs.In this paper, we propose a novel DivCo framework to properly constrain both "positive" and "negative" relations between the generated images specified in the latent space. To the best of our knowledge, this is the first attempt to use contrastive learning for diverse conditional image synthesis. A novel latent-augmented contrastive loss is introduced, which encourages images generated from adjacent latent codes to be similar and those generated from distinct latent codes to be dissimilar. The proposed latent-augmented contrastive loss is well compatible with various cGAN architectures. Extensive experiments demonstrate that the proposed DivCo can produce more diverse images than state-of-the-art methods without sacrificing visual quality in multiple unpaired and paired image generation tasks. Training code and pretrained models are available at https://github.com/ruiliu-ai/DivCo.

Graph-neural-networks (GNN) is a rising trend for fewshot learning. A critical component in GNN is the affinity. Typically, affinity in GNN is mainly computed in the feature space, e.g., pairwise features, and does not take fully advantage of semantic labels associated to these features. In this paper, we propose a novel Mutual CRF-GNN (MCGN). In this MCGN, the labels and features of support data are used by the CRF for inferring GNN affinities in a principled and probabilistic way. Specifically, we construct a Conditional Random Field (CRF) conditioned on labels and features of support data to infer a affinity in the label space. Such affinity is fed to the GNN as the node-wise affinity. GNN and CRF mutually contributes to each other in MCGN. For GNN, CRF provides valuable affinity information. For CRF, GNN provides better features for inferring affinity. Experimental results show that our approach outperforms stateof-the-arts on datasets miniImageNet, tieredImageNet, and CIFAR-FS on both 5-way 1-shot and 5-way 5-shot settings.

Unsupervised object re-identification targets at learning discriminative representations for object retrieval without any annotations. Clustering-based methods [27], [46], [10] conduct training with the generated pseudo labels and currently dominate this research direction. However, they still suffer from the issue of pseudo label noise. To tackle the challenge, we propose to properly estimate pseudo label similarities between consecutive training generations with clustering consensus and refine pseudo labels with temporally propagated and ensembled pseudo labels. To the best of our knowledge, this is the first attempt to leverage the spirit of temporal ensembling [25] to improve classification with dynamically changing classes over generations. The proposed pseudo label refinery strategy is simple yet effective and can be seamlessly integrated into existing clustering-based unsupervised re-identification methods. With our proposed approach, state-of-the-art method [10] can be further boosted with up to 8.8% mAP improvements on the challenging MSMT17 [39] dataset.

Correspondence pruning aims to correctly remove false matches (outliers) from an initial set of putative correspondences. The pruning process is challenging since putative matches are typically extremely unbalanced, largely dominated by outliers, and the random distribution of such outliers further complicates the learning process for learning-based methods. To address this issue, we propose to progressively prune the correspondences via a local-to-global consensus learning procedure. We introduce a "pruning" block that lets us identify reliable candidates among the initial matches according to consensus scores estimated using local-to-global dynamic graphs. We then achieve progressive pruning by stacking multiple pruning blocks sequentially. Our method outperforms state-of-the-arts on robust line fitting, camera pose estimation and retrieval-based image localization benchmarks by significant margins and shows promising generalization ability to different datasets and detector/descriptor combinations.

Adaptive person re-identification (adaptive ReID) targets at transferring learned knowledge from the labeled source domain to the unlabeled target domain. Pseudo-label-based methods that alternatively generate pseudo labels and optimize the training model have demonstrated great effectiveness in this field. However, the generated pseudo labels are inaccurate and cannot reflect the true semantic meaning of the unlabeled samples. We consider such inaccuracy stems from both the lagged update of the pseudo labels as well as the simple criterion of the employed clustering method. To tackle the problem, we propose an online pseudo label generation by hierarchical cluster dynamics for adaptive ReID. In particular, hierarchical label banks are constructed for all the samples in the dataset, and we update the pseudo labels of the sample in each coming mini-batch, performing the model optimization and the label generation simultaneously. A new hierarchical cluster dynamics is built for the label update, where cluster merge and cluster split are driven by a possibility computed by the label propagation. Our method can achieve better pseudo labels and higher reid accuracy. Extensive experiments on Market-to-Duke, Duke-to-Market, MSMT-to-Market, MSMT-to-Duke, Market-to-MSMT, and Duke-to-MSMT verify the effectiveness of our proposed method.

Domain adaptive object re-ID aims to transfer the learned knowledge from the labeled source domain to the unlabeled target domain to tackle the open-class re-identification problems. Although state-of-the-art pseudo-label-based methods have achieved great success, they did not make full use of all valuable information because of the domain gap and unsatisfying clustering performance. To solve these problems, we propose a novel self-paced contrastive learning framework with hybrid memory. The hybrid memory dynamically generates source-domain class-level, target-domain cluster-level and un-clustered instance-level supervisory signals for learning feature representations. Different from the conventional contrastive learning strategy, the proposed framework jointly distinguishes source-domain classes, and target-domain clusters and un-clustered instances. Most importantly, the proposed self-paced method gradually creates more reliable clusters to refine the hybrid memory and learning targets, and is shown to be the key to our outstanding performance. Our method outperforms state-of-the-arts on multiple domain adaptation tasks of object re-ID and even boosts the performance on the source domain without any extra annotations. Our generalized version on unsupervised object re-ID surpasses state-of-the-art algorithms by considerable 16.7% and 7.9% on Market-1501 and MSMT17 benchmarks.

Person re-identification (reID) is an important task that requires to retrieve a person's images from an image dataset, given one image of the person of interest. For learning robust person features, the pose variation of person images is one of the key challenges. Existing works targeting the problem either perform human alignment, or learn human-region-based representations. Extra pose information and computational cost is generally required for inference. To solve this issue, a Feature Distilling Generative Adversarial Network (FD-GAN) is proposed for learning identity-related and pose-unrelated representations. It is a novel framework based on a Siamese structure with multiple novel discriminators on human poses and identities. In addition to the discriminators, a novel same-pose loss is also integrated, which requires appearance of a same person's generated images to be similar. After learning pose-unrelated person features with pose guidance, no auxiliary pose information and additional computational cost is required during testing. Our proposed FD-GAN achieves state-of-the-art performance on three person reID datasets, which demonstrates that the effectiveness and robust feature distilling capability of the proposed FD-GAN.