{"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/clipcap-clip-prefix-for-image-captioning","title":"ClipCap: CLIP Prefix for Image Captioning","arxiv_id":"2111.09734","date":"2021-11-18","proceeding":null,"authors":["Ron Mokady","Amir Hertz","Amit H. Bermano"],"abstract":"Image captioning is a fundamental task in vision-language understanding, where the model predicts a textual informative caption to a given input image. In this paper, we present a simple approach to address this task. We use CLIP encoding as a prefix to the caption, by employing a simple mapping network, and then fine-tunes a language model to generate the image captions. The recently proposed CLIP model contains rich semantic features which were trained with textual context, making it best for vision-language perception. Our key idea is that together with a pre-trained language model (GPT2), we obtain a wide understanding of both visual and textual data. Hence, our approach only requires rather quick training to produce a competent captioning model. Without additional annotations or pre-training, it efficiently generates meaningful captions for large-scale and diverse datasets. Surprisingly, our method works well even when only the mapping network is trained, while both CLIP and the language model remain frozen, allowing a lighter architecture with less trainable parameters. Through quantitative evaluation, we demonstrate our model achieves comparable results to state-of-the-art methods on the challenging Conceptual Captions and nocaps datasets, while it is simpler, faster, and lighter. Our code is available in https://github.com/rmokady/CLIP_prefix_caption.","url_abs":"https://arxiv.org/abs/2111.09734v1","url_pdf":"https://arxiv.org/pdf/2111.09734v1.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":"clipcap-clip-prefix-for-image-captioning","repo_url":"https://github.com/rmokady/clip_prefix_caption","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"clipcap-clip-prefix-for-image-captioning","repo_url":"https://github.com/Japanese-Image-Captioning/ClipCap-for-Japanese","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"clipcap-clip-prefix-for-image-captioning","repo_url":"https://github.com/sithu31296/image-captioning","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"clipcap-clip-prefix-for-image-captioning","repo_url":"https://github.com/MS-P3/code7/tree/main/x_clip","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null}],"tasks":[{"task_slug":"image-captioning","task_name":"Image Captioning"},{"task_slug":"language-modeling","task_name":"Language Modeling"},{"task_slug":"language-modelling","task_name":"Language Modelling"}],"methods":[{"method_slug":"adam","method_name":"Adam"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"attention-dropout","method_name":"Attention Dropout"},{"method_slug":"bpe","method_name":"BPE"},{"method_slug":"clip","method_name":"CLIP"},{"method_slug":"cosine-annealing","method_name":"Cosine Annealing"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"discriminative-fine-tuning","method_name":"Discriminative Fine-Tuning"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"gpt-2","method_name":"GPT-2"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"linear-warmup-with-cosine-annealing","method_name":"Linear Warmup With Cosine Annealing"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"weight-decay","method_name":"Weight Decay"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/image-captioning-on-coco-captions","task":"Image Captioning","dataset":"COCO Captions","model":"ClipCap (Transformer)","rank_in_archive_order":30,"of":41,"metrics":{"BLEU-4":"33.53","CIDER":"113.08","METEOR":"27.45","SPICE":"21.05"},"uses_additional_data":false},{"leaderboard":"/sota/image-captioning-on-coco-captions","task":"Image Captioning","dataset":"COCO Captions","model":"ClipCap (MLP + GPT2 tuning)","rank_in_archive_order":31,"of":41,"metrics":{"BLEU-4":"32.15","CIDER":"108.35","METEOR":"27.1","SPICE":"20.12"},"uses_additional_data":false},{"leaderboard":"/sota/image-captioning-on-conceptual-captions","task":"Image Captioning","dataset":"Conceptual Captions","model":"ClipCap (MLP + GPT2 tuning)","rank_in_archive_order":1,"of":2,"metrics":{"CIDEr":"87.26","ROUGE-L":"26.71","SPICE":"18.5"},"uses_additional_data":false},{"leaderboard":"/sota/image-captioning-on-conceptual-captions","task":"Image Captioning","dataset":"Conceptual Captions","model":"ClipCap (Transformer)","rank_in_archive_order":2,"of":2,"metrics":{"CIDEr":"71.82","ROUGE-L":"25.12","SPICE":"16.07"},"uses_additional_data":false},{"leaderboard":"/sota/image-captioning-on-nocaps-entire","task":"Image Captioning","dataset":"nocaps entire","model":"ClipCap (Transformer)","rank_in_archive_order":26,"of":39,"metrics":{"CIDEr":"65.83","SPICE":"10.86"},"uses_additional_data":false},{"leaderboard":"/sota/image-captioning-on-nocaps-entire","task":"Image Captioning","dataset":"nocaps entire","model":"ClipCap (MLP + GPT2 tuning)","rank_in_archive_order":27,"of":39,"metrics":{"CIDEr":"65.7","SPICE":"11.1"},"uses_additional_data":false},{"leaderboard":"/sota/image-captioning-on-nocaps-in-domain","task":"Image Captioning","dataset":"nocaps in-domain","model":"ClipCap (Transformer)","rank_in_archive_order":23,"of":41,"metrics":{"CIDEr":"84.85","SPICE":"12.14"},"uses_additional_data":false},{"leaderboard":"/sota/image-captioning-on-nocaps-in-domain","task":"Image Captioning","dataset":"nocaps in-domain","model":"ClipCap (MLP + GPT2 tuning)","rank_in_archive_order":28,"of":41,"metrics":{"CIDEr":"79.73","SPICE":"12.2"},"uses_additional_data":false},{"leaderboard":"/sota/image-captioning-on-nocaps-near-domain","task":"Image Captioning","dataset":"nocaps near-domain","model":"ClipCap (MLP + GPT2 tuning)","rank_in_archive_order":26,"of":40,"metrics":{"CIDEr":"67.69","SPICE":"11.26"},"uses_additional_data":false},{"leaderboard":"/sota/image-captioning-on-nocaps-near-domain","task":"Image Captioning","dataset":"nocaps near-domain","model":"ClipCap (Transformer)","rank_in_archive_order":27,"of":40,"metrics":{"CIDEr":"66.82","SPICE":"10.92"},"uses_additional_data":false},{"leaderboard":"/sota/image-captioning-on-nocaps-out-of-domain","task":"Image Captioning","dataset":"nocaps out-of-domain","model":"ClipCap (MLP + GPT2 tuning)","rank_in_archive_order":28,"of":40,"metrics":{"CIDEr":"49.35","SPICE":"9.7"},"uses_additional_data":false},{"leaderboard":"/sota/image-captioning-on-nocaps-out-of-domain","task":"Image Captioning","dataset":"nocaps out-of-domain","model":"ClipCap (Transformer)","rank_in_archive_order":29,"of":40,"metrics":{"CIDEr":"49.14","SPICE":"9.57"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2111.09734","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2111.09734"}},"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/MS-P3/code7/tree/main/x_clip","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/Japanese-Image-Captioning/ClipCap-for-Japanese","reach":{"status":"ok","spdx":"MIT"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/rmokady/clip_prefix_caption","reach":{"status":"ok","spdx":"MIT"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/sithu31296/image-captioning","reach":null}],"summary":{"ran":2,"ran_draft_wrong":1,"unverified":3},"by_repo_kind":{"official":{"samples":5,"ran":2,"repositories":1},"listed":{"samples":1,"ran":1,"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":"e33b112514a9e4b0","entry":"generate2","repo":"rmokady/clip_prefix_caption","repo_kind":"official","path":"predict.py","file_url":"https://github.com/rmokady/clip_prefix_caption/blob/HEAD/predict.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"e33b112514a9e4b0"}},{"code_sha256_prefix":"883320eb62f15d12","entry":"generate_beam","repo":"rmokady/clip_prefix_caption","repo_kind":"official","path":"predict.py","file_url":"https://github.com/rmokady/clip_prefix_caption/blob/HEAD/predict.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"883320eb62f15d12"}},{"code_sha256_prefix":"6800239cecf973e8","entry":"get_annotation_file","repo":"sithu31296/image-captioning","repo_kind":"listed","path":"tools/val.py","file_url":"https://github.com/sithu31296/image-captioning/blob/HEAD/tools/val.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":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"6800239cecf973e8"}},{"code_sha256_prefix":"c68cb2401825d6a1","entry":"add_period","repo":"rmokady/clip_prefix_caption","repo_kind":"official","path":"parse_conceptual.py","file_url":"https://github.com/rmokady/clip_prefix_caption/blob/HEAD/parse_conceptual.py","link_basis":"harvester_set","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":"c68cb2401825d6a1"}},{"code_sha256_prefix":"0a286fb23ba005b9","entry":"get_image","repo":"rmokady/clip_prefix_caption","repo_kind":"official","path":"parse_conceptual.py","file_url":"https://github.com/rmokady/clip_prefix_caption/blob/HEAD/parse_conceptual.py","link_basis":"harvester_set","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":"0a286fb23ba005b9"}},{"code_sha256_prefix":"50449115c68a0ddb","entry":"thread","repo":"rmokady/clip_prefix_caption","repo_kind":"official","path":"parse_conceptual.py","file_url":"https://github.com/rmokady/clip_prefix_caption/blob/HEAD/parse_conceptual.py","link_basis":"harvester_set","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":"50449115c68a0ddb"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}