{"url":"/method/moco-v3","slug":"moco-v3","name":"MoCo v3","full_name":"MoCo v3","full_name_withheld":false,"description_markdown":"**MoCo v3** aims to stabilize training of self-supervised ViTs. MoCo v3 is an incremental improvement of MoCo v1/2. Two crops are used for each image under random data augmentation. They are encoded by two encoders $f_q$ and $f_k$ with output vectors $q$ and $k$. $q$ behaves like a \"query\", where the goal of learning is to retrieve the corresponding \"key\". The objective is to minimize a contrastive loss function of the following form: \r\n\r\n$$\r\n\\mathcal{L_q}=-\\log \\frac{\\exp \\left(q \\cdot k^{+} / \\tau\\right)}{\\exp \\left(q \\cdot k^{+} / \\tau\\right)+\\sum_{k^{-}} \\exp \\left(q \\cdot k^{-} / \\tau\\right)}\r\n$$\r\n\r\nThis approach aims to train the Transformer in the contrastive/Siamese paradigm. The encoder $f_q$ consists of a backbone (e.g., ResNet and ViT), a projection head, and an extra prediction head. The encoder $f_k$ has the back the backbone and projection head but not the prediction head. $f_k$ is updated by the moving average of $f_q$, excluding the prediction head.","description_state":"present","introduced_year":null,"introduced_by":{"title":"An Empirical Study of Training Self-Supervised Vision Transformers","paper":"/paper/an-empirical-study-of-training-self","first_author":"Xinlei Chen","n_authors":3,"url_abs":null,"archive_paper_url":"https://paperswithcode.com/paper/an-empirical-study-of-training-self"},"source":{"url":"https://arxiv.org/abs/2104.02057v4","title":"An Empirical Study of Training Self-Supervised Vision Transformers","url_on_a_paper_host":true},"code_snippet_url":null,"code_snippet_url_on_a_code_host":false,"categories":[{"area":"Computer Vision","area_id":"computer-vision","collection":"Vision Transformers","url":"/methods/category/vision-transformers","pwc_aliases":["vision-transformer"]}],"n_papers_tagged":10,"archive_num_papers":10,"papers_newest_first":[{"paper":null,"title":"Task-Specific Knowledge Distillation from the Vision Foundation Model for Enhanced Medical Image Segmentation","date":"2025-03-10","arxiv_id":"2503.06976","n_code_links":0,"syntology":null},{"paper":"/paper/enhancing-contrastive-learning-inspired-by","title":"Enhancing Contrastive Learning Inspired by the Philosophy of \"The Blind Men and the Elephant\"","date":"2024-12-21","arxiv_id":"2412.16522","n_code_links":1,"syntology":null},{"paper":null,"title":"SRA: A Novel Method to Improve Feature Embedding in Self-supervised Learning for Histopathological Images","date":"2024-10-23","arxiv_id":"2410.17514","n_code_links":0,"syntology":null},{"paper":"/paper/improving-visual-prompt-tuning-for-self","title":"Improving Visual Prompt Tuning for Self-supervised Vision Transformers","date":"2023-06-08","arxiv_id":"2306.05067","n_code_links":1,"syntology":{"ran":4,"of":4,"unverified":0,"pointer_only":0}},{"paper":"/paper/internet-explorer-targeted-representation","title":"Internet Explorer: Targeted Representation Learning on the Open Web","date":"2023-02-27","arxiv_id":"2302.14051","n_code_links":1,"syntology":null},{"paper":"/paper/pretraining-the-vision-transformer-using-self","title":"Pretraining the Vision Transformer using self-supervised methods for vision based Deep Reinforcement Learning","date":"2022-09-22","arxiv_id":"2209.10901","n_code_links":1,"syntology":null},{"paper":"/paper/fast-moco-boost-momentum-based-contrastive","title":"Fast-MoCo: Boost Momentum-based Contrastive Learning with Combinatorial Patches","date":"2022-07-17","arxiv_id":"2207.08220","n_code_links":1,"syntology":{"ran":1,"of":4,"unverified":3,"pointer_only":4}},{"paper":"/paper/unifying-architectures-tasks-and-modalities","title":"OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework","date":"2022-02-07","arxiv_id":"2202.03052","n_code_links":4,"syntology":{"ran":1,"of":1,"unverified":0,"pointer_only":0}},{"paper":"/paper/self-supervised-learning-with-swin","title":"Self-Supervised Learning with Swin Transformers","date":"2021-05-10","arxiv_id":"2105.04553","n_code_links":6,"syntology":null},{"paper":"/paper/an-empirical-study-of-training-self","title":"An Empirical Study of Training Self-Supervised Vision Transformers","date":"2021-04-05","arxiv_id":"2104.02057","n_code_links":9,"syntology":{"ran":2,"of":2,"unverified":0,"pointer_only":2}}],"papers_shown":10,"tasks":[{"task":"/task/self-supervised-learning","name":"Self-Supervised Learning","papers":6},{"task":"/task/contrastive-learning","name":"Contrastive Learning","papers":3},{"task":"/task/self-supervised-image-classification","name":"Self-Supervised Image Classification","papers":3},{"task":"/task/semantic-segmentation","name":"Semantic Segmentation","papers":3},{"task":"/task/image-classification","name":"Image Classification","papers":2},{"task":"/task/linear-evaluation","name":"Linear evaluation","papers":2},{"task":"/task/representation-learning","name":"Representation Learning","papers":2},{"task":"/task/image-classification","name":"image-classification","papers":2},{"task":"/task/atari-games","name":"Atari Games","papers":1},{"task":"/task/atari-games-100k","name":"Atari Games 100k","papers":1},{"task":"/task/classification-1","name":"Classification","papers":1},{"task":"/task/data-augmentation","name":"Data Augmentation","papers":1},{"task":"/task/deep-reinforcement-learning","name":"Deep Reinforcement Learning","papers":1},{"task":null,"name":"GPU","papers":1},{"task":"/task/image-augmentation","name":"Image Augmentation","papers":1},{"task":"/task/image-captioning","name":"Image Captioning","papers":1},{"task":"/task/image-generation","name":"Image Generation","papers":1},{"task":"/task/image-segmentation","name":"Image Segmentation","papers":1},{"task":"/task/knowledge-distillation","name":"Knowledge Distillation","papers":1},{"task":"/task/language-modeling","name":"Language Modeling","papers":1}],"tasks_shown":20,"n_tasks":42,"usage_by_year":[{"year":"2021","papers":2},{"year":"2022","papers":3},{"year":"2023","papers":2},{"year":"2024","papers":2},{"year":"2025","papers":1}],"row_source":"methods_table","archive":{"source":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","archive_url":"https://paperswithcode.com/method/moco-v3"},"syntology_read_at":"2026-09-24T18:15:14+00:00"}