{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/vit-lens-towards-omni-modal-representations","title":"ViT-Lens: Initiating Omni-Modal Exploration through 3D Insights","arxiv_id":"2308.10185","date":"2023-08-20","proceeding":null,"authors":["Weixian Lei","Yixiao Ge","Jianfeng Zhang","Dylan Sun","Kun Yi","Ying Shan","Mike Zheng Shou"],"abstract":"Though the success of CLIP-based training recipes in vision-language models, their scalability to more modalities (e.g., 3D, audio, etc.) is limited to large-scale data, which is expensive or even inapplicable 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 to a pre-defined space. Specifically, the modality-specific lens is tuned to project multimodal signals to the shared embedding space, which are then processed by a strong ViT that carries pre-trained image knowledge. The encoded multimodal representations are optimized toward aligning with the modal-independent space, pre-defined by off-the-shelf foundation models. A well-trained lens with a ViT backbone has the potential to serve as one of these foundation models, supervising the learning of subsequent modalities. ViT-Lens provides a unified solution for representation learning of increasing modalities with two appealing benefits: (i) Exploiting the pretrained ViT across tasks and domains effectively with efficient data regime; (ii) Emergent downstream capabilities of novel modalities are demonstrated due to the modality alignment space. We evaluate ViT-Lens in the context of 3D as an initial verification. In zero-shot 3D classification, ViT-Lens achieves substantial improvements over previous state-of-the-art, showing 52.0% accuracy on Objaverse-LVIS, 87.4% on ModelNet40, and 60.6% on ScanObjectNN. Furthermore, we enable zero-shot 3D question-answering by simply integrating the trained 3D lens into the InstructBLIP model without any adaptation. We will release the results of ViT-Lens on more modalities in the near future.","url_abs":"https://arxiv.org/abs/2308.10185v2","url_pdf":"https://arxiv.org/pdf/2308.10185v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"vit-lens-towards-omni-modal-representations","repo_url":"https://github.com/TencentARC/ViT-Lens","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"3d-classification","task_name":"3D Classification"},{"task_slug":"question-answering","task_name":"Question Answering"},{"task_slug":"representation-learning","task_name":"Representation Learning"},{"task_slug":"training-free-3d-point-cloud-classification","task_name":"Training-free 3D Point Cloud Classification"},{"task_slug":"zero-shot-transfer-3d-point-cloud","task_name":"Zero-Shot Transfer 3D Point Cloud Classification"},{"task_slug":"zero-shot-3d-classification","task_name":"Zero-shot 3D classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/zero-shot-transfer-3d-point-cloud","task":"Zero-Shot Transfer 3D Point Cloud Classification","dataset":"ModelNet40","model":"ViT-Lens","rank_in_archive_order":2,"of":16,"metrics":{"Accuracy (%)":"87.6"},"uses_additional_data":true},{"leaderboard":"/sota/zero-shot-transfer-3d-point-cloud-2","task":"Zero-Shot Transfer 3D Point Cloud Classification","dataset":"ScanObjectNN","model":"ViT-Lens","rank_in_archive_order":4,"of":10,"metrics":{"OBJ_ONLY Accuracy(%)":"60.1"},"uses_additional_data":true}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2308.10185","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}