{"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/internimage-exploring-large-scale-vision","title":"InternImage: Exploring Large-Scale Vision Foundation Models with Deformable Convolutions","arxiv_id":"2211.05778","date":"2022-11-10","proceeding":"CVPR 2023 1","authors":["Wenhai Wang","Jifeng Dai","Zhe Chen","Zhenhang Huang","Zhiqi Li","Xizhou Zhu","Xiaowei Hu","Tong Lu","Lewei Lu","Hongsheng Li","Xiaogang Wang","Yu Qiao"],"abstract":"Compared to the great progress of large-scale vision transformers (ViTs) in recent years, large-scale models based on convolutional neural networks (CNNs) are still in an early state. This work presents a new large-scale CNN-based foundation model, termed InternImage, which can obtain the gain from increasing parameters and training data like ViTs. Different from the recent CNNs that focus on large dense kernels, InternImage takes deformable convolution as the core operator, so that our model not only has the large effective receptive field required for downstream tasks such as detection and segmentation, but also has the adaptive spatial aggregation conditioned by input and task information. As a result, the proposed InternImage reduces the strict inductive bias of traditional CNNs and makes it possible to learn stronger and more robust patterns with large-scale parameters from massive data like ViTs. The effectiveness of our model is proven on challenging benchmarks including ImageNet, COCO, and ADE20K. It is worth mentioning that InternImage-H achieved a new record 65.4 mAP on COCO test-dev and 62.9 mIoU on ADE20K, outperforming current leading CNNs and ViTs. The code will be released at https://github.com/OpenGVLab/InternImage.","url_abs":"https://arxiv.org/abs/2211.05778v4","url_pdf":"https://arxiv.org/pdf/2211.05778v4.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":"internimage-exploring-large-scale-vision","repo_url":"https://github.com/opengvlab/internimage","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"internimage-exploring-large-scale-vision","repo_url":"https://github.com/OpenGVLab/M3I-Pretraining","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"internimage-exploring-large-scale-vision","repo_url":"https://github.com/chenller/mmseg-extension","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"2d-object-detection","task_name":"2D Object Detection"},{"task_slug":"classification-1","task_name":"Classification"},{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"instance-segmentation","task_name":"Instance Segmentation"},{"task_slug":"object-detection","task_name":"Object Detection"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"deformable-convolution","method_name":"Deformable Convolution"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/2d-object-detection-on-bdd100k-val","task":"2D Object Detection","dataset":"BDD100K val","model":"InternImage-H","rank_in_archive_order":1,"of":1,"metrics":{"mAP":"38.8"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-imagenet","task":"Image Classification","dataset":"ImageNet","model":"InternImage-H","rank_in_archive_order":16,"of":1060,"metrics":{"GFLOPs":"1478","Number of params":"1080M","Top 1 Accuracy":"89.6%"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-imagenet","task":"Image Classification","dataset":"ImageNet","model":"InternImage-XL","rank_in_archive_order":64,"of":1060,"metrics":{"GFLOPs":"163","Number of params":"335M","Top 1 Accuracy":"88%"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-imagenet","task":"Image Classification","dataset":"ImageNet","model":"InternImage-L","rank_in_archive_order":77,"of":1060,"metrics":{"GFLOPs":"108","Number of params":"223M","Top 1 Accuracy":"87.7%"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-imagenet","task":"Image Classification","dataset":"ImageNet","model":"InternImage-B","rank_in_archive_order":281,"of":1060,"metrics":{"GFLOPs":"16","Number of params":"97M","Top 1 Accuracy":"84.9%"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-imagenet","task":"Image Classification","dataset":"ImageNet","model":"InternImage-S","rank_in_archive_order":338,"of":1060,"metrics":{"GFLOPs":"8","Number of params":"50M","Top 1 Accuracy":"84.2%"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-imagenet","task":"Image Classification","dataset":"ImageNet","model":"InternImage-DCNv3-G (M3I Pre-training)","rank_in_archive_order":1060,"of":1060,"metrics":{"Number of params":"3000M"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-places205","task":"Image Classification","dataset":"Places205","model":"InternImage-H","rank_in_archive_order":1,"of":15,"metrics":{"Top 1 Accuracy":"71.7%"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-places365","task":"Image Classification","dataset":"Places365","model":"InternImage-H（CNN）","rank_in_archive_order":3,"of":7,"metrics":{"Top 1 Accuracy":"61.2%"},"uses_additional_data":true},{"leaderboard":"/sota/image-classification-on-inaturalist-2018","task":"Image Classification","dataset":"iNaturalist 2018","model":"InternImage-H","rank_in_archive_order":3,"of":60,"metrics":{"Top-1 Accuracy":"92.6%"},"uses_additional_data":true},{"leaderboard":"/sota/instance-segmentation-on-coco-minival","task":"Instance Segmentation","dataset":"COCO minival","model":"InternImage-H","rank_in_archive_order":3,"of":93,"metrics":{"AP50":"80.1","AP75":"61.5","APL":"74.4","APM":"58.4","APS":"37.9","mask AP":"55.4"},"uses_additional_data":true},{"leaderboard":"/sota/instance-segmentation-on-coco-minival","task":"Instance Segmentation","dataset":"COCO minival","model":"InternImage-XL","rank_in_archive_order":32,"of":93,"metrics":{"GFLOPs":"1782","Params (M)":"387","mask AP":"48.8"},"uses_additional_data":false},{"leaderboard":"/sota/instance-segmentation-on-coco-minival","task":"Instance Segmentation","dataset":"COCO minival","model":"InternImage-L","rank_in_archive_order":35,"of":93,"metrics":{"GFLOPs":"1399","Params (M)":"277","box AP":"56.1","mask AP":"48.5"},"uses_additional_data":false},{"leaderboard":"/sota/instance-segmentation-on-coco-minival","task":"Instance Segmentation","dataset":"COCO minival","model":"InternImage-S","rank_in_archive_order":52,"of":93,"metrics":{"GFLOPs":"340","Params (M)":"69","box AP":"49.7","mask AP":"44.5"},"uses_additional_data":false},{"leaderboard":"/sota/instance-segmentation-on-coco-minival","task":"Instance Segmentation","dataset":"COCO minival","model":"InternImage-T","rank_in_archive_order":58,"of":93,"metrics":{"GFLOPs":"270","Params (M)":"49","box AP":"49.1","mask AP":"43.7"},"uses_additional_data":false},{"leaderboard":"/sota/instance-segmentation-on-coco-minival","task":"Instance Segmentation","dataset":"COCO minival","model":"InternImage-B","rank_in_archive_order":93,"of":93,"metrics":{"GFLOPs":"501","Params (M)":"115"},"uses_additional_data":false},{"leaderboard":"/sota/instance-segmentation-on-coco","task":"Instance Segmentation","dataset":"COCO test-dev","model":"InternImage-H","rank_in_archive_order":107,"of":112,"metrics":{"AP50":"80.8","AP75":"62.2","APL":"70.3","APM":"58.9","APS":"41.0"},"uses_additional_data":true},{"leaderboard":"/sota/object-detection-on-coco-minival","task":"Object Detection","dataset":"COCO minival","model":"InternImage-H","rank_in_archive_order":4,"of":220,"metrics":{"box AP":"65.0"},"uses_additional_data":true},{"leaderboard":"/sota/object-detection-on-coco-minival","task":"Object Detection","dataset":"COCO minival","model":"InternImage-XL","rank_in_archive_order":10,"of":220,"metrics":{"box AP":"64.2"},"uses_additional_data":true},{"leaderboard":"/sota/object-detection-on-coco","task":"Object Detection","dataset":"COCO test-dev","model":"InternImage-H (M3I Pre-training)","rank_in_archive_order":2,"of":225,"metrics":{"Params (M)":"2180","box mAP":"65.5"},"uses_additional_data":false},{"leaderboard":"/sota/object-detection-on-coco","task":"Object Detection","dataset":"COCO test-dev","model":"InternImage-XL","rank_in_archive_order":10,"of":225,"metrics":{"Params (M)":"602","box mAP":"64.3"},"uses_additional_data":false},{"leaderboard":"/sota/object-detection-on-coco-o","task":"Object Detection","dataset":"COCO-O","model":"InternImage-L (Cascade Mask R-CNN)","rank_in_archive_order":8,"of":45,"metrics":{"Average mAP":"37.0","Effective Robustness":"11.72"},"uses_additional_data":false},{"leaderboard":"/sota/object-detection-on-crowdhuman-full-body","task":"Object Detection","dataset":"CrowdHuman (full body)","model":"InternImage-H","rank_in_archive_order":1,"of":19,"metrics":{"AP":"97.2"},"uses_additional_data":false},{"leaderboard":"/sota/object-detection-on-lvis-v1-0-minival","task":"Object Detection","dataset":"LVIS v1.0 minival","model":"InternImage-H","rank_in_archive_order":5,"of":6,"metrics":{"box AP":"65.8"},"uses_additional_data":false},{"leaderboard":"/sota/object-detection-on-lvis-v1-0-val","task":"Object Detection","dataset":"LVIS v1.0 val","model":"InternImage-H","rank_in_archive_order":3,"of":15,"metrics":{"box AP":"63.2"},"uses_additional_data":true},{"leaderboard":"/sota/object-detection-on-openimages-v6","task":"Object Detection","dataset":"OpenImages-v6","model":"InternImage-H","rank_in_archive_order":2,"of":2,"metrics":{"box AP":"74.1"},"uses_additional_data":false},{"leaderboard":"/sota/object-detection-on-pascal-voc-2012","task":"Object Detection","dataset":"PASCAL VOC 2012","model":"InternImage-H","rank_in_archive_order":1,"of":7,"metrics":{"MAP":"97.2"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-segmentation-on-ade20k","task":"Semantic Segmentation","dataset":"ADE20K","model":"InternImage-H","rank_in_archive_order":3,"of":235,"metrics":{"GFLOPs":"4635","Params (M)":"1310","Validation mIoU":"62.9"},"uses_additional_data":true},{"leaderboard":"/sota/semantic-segmentation-on-ade20k","task":"Semantic Segmentation","dataset":"ADE20K","model":"InternImage-XL","rank_in_archive_order":48,"of":235,"metrics":{"GFLOPs":"3142","Params (M)":"368","Validation mIoU":"55.3"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-segmentation-on-ade20k","task":"Semantic Segmentation","dataset":"ADE20K","model":"InternImage-L","rank_in_archive_order":67,"of":235,"metrics":{"GFLOPs":"2526","Params (M)":"256","Validation mIoU":"54.1"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-segmentation-on-ade20k","task":"Semantic Segmentation","dataset":"ADE20K","model":"InternImage-B","rank_in_archive_order":98,"of":235,"metrics":{"GFLOPs":"1185","Params (M)":"128","Validation mIoU":"51.3"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-segmentation-on-ade20k","task":"Semantic Segmentation","dataset":"ADE20K","model":"InternImage-S","rank_in_archive_order":107,"of":235,"metrics":{"GFLOPs":"1017","Params (M)":"80","Validation mIoU":"50.9"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-segmentation-on-ade20k","task":"Semantic Segmentation","dataset":"ADE20K","model":"InternImage-T","rank_in_archive_order":150,"of":235,"metrics":{"GFLOPs":"944","Params (M)":"59","Validation mIoU":"48.1"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-segmentation-on-ade20k","task":"Semantic Segmentation","dataset":"ADE20K","model":"InternImage-H (M3I Pre-training)","rank_in_archive_order":231,"of":235,"metrics":{"Params (M)":"1310"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-segmentation-on-cityscapes","task":"Semantic Segmentation","dataset":"Cityscapes test","model":"InternImage-H","rank_in_archive_order":3,"of":105,"metrics":{"Mean IoU (class)":"86.1%"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-segmentation-on-cityscapes-val","task":"Semantic Segmentation","dataset":"Cityscapes val","model":"InternImage-H","rank_in_archive_order":4,"of":99,"metrics":{"mIoU":"87"},"uses_additional_data":true},{"leaderboard":"/sota/semantic-segmentation-on-cityscapes-val","task":"Semantic Segmentation","dataset":"Cityscapes val","model":"InternImage-XL","rank_in_archive_order":6,"of":99,"metrics":{"mIoU":"86.4"},"uses_additional_data":true},{"leaderboard":"/sota/semantic-segmentation-on-pascal-context","task":"Semantic Segmentation","dataset":"PASCAL Context","model":"InternImage-H","rank_in_archive_order":3,"of":66,"metrics":{"mIoU":"70.3"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-segmentation-on-replica","task":"Semantic Segmentation","dataset":"Replica","model":"InternImage","rank_in_archive_order":2,"of":5,"metrics":{"mIoU":"38.4"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2211.05778","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2211.05778"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/opengvlab/internimage","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/OpenGVLab/M3I-Pretraining","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/chenller/mmseg-extension","reach":null}],"summary":{"ran_draft_wrong":2,"unverified":2},"by_repo_kind":{"official":{"samples":2,"ran":0,"repositories":1},"listed":{"samples":2,"ran":2,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"f5dee2f80ed7988a","entry":"build_act_layer","repo":"chenller/mmseg-extension","repo_kind":"listed","path":"mmsegext/models/backbones/intern_image.py","file_url":"https://github.com/chenller/mmseg-extension/blob/HEAD/mmsegext/models/backbones/intern_image.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"f5dee2f80ed7988a"}},{"code_sha256_prefix":"e7f1a63c2e99c70f","entry":"build_norm_layer","repo":"chenller/mmseg-extension","repo_kind":"listed","path":"mmsegext/models/backbones/intern_image.py","file_url":"https://github.com/chenller/mmseg-extension/blob/HEAD/mmsegext/models/backbones/intern_image.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"e7f1a63c2e99c70f"}},{"code_sha256_prefix":"5f9bbcaceee97948","entry":"main","repo":"opengvlab/internimage","repo_kind":"official","path":"classification/extract_feature.py","file_url":"https://github.com/opengvlab/internimage/blob/HEAD/classification/extract_feature.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"5f9bbcaceee97948"}},{"code_sha256_prefix":"44925cf3831bfe79","entry":"rgetattr","repo":"opengvlab/internimage","repo_kind":"official","path":"classification/extract_feature.py","file_url":"https://github.com/opengvlab/internimage/blob/HEAD/classification/extract_feature.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"44925cf3831bfe79"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}