{"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/generative-pretraining-from-pixels","title":"Generative Pretraining from Pixels","arxiv_id":null,"date":"2020-07-17","proceeding":"ICML 2020 1","authors":["Mark Chen","Alec Radford","Rewon Child","Jeff Wu","Heewoo Jun","Prafulla Dhariwal","David Luan","Ilya Sutskever"],"abstract":"Inspired by progress in unsupervised representation learning for natural language, we examine whether similar models can learn useful representations for images. We train a sequence Transformer to auto-regressively predict pixels, without incorporating knowledge of the 2D input structure. Despite training on low-resolution ImageNet without labels, we find that a GPT-2 scale model learns strong image representations as measured by linear probing, fine-tuning, and low-data classification. On CIFAR-10, we achieve 96.3% accuracy with a linear probe, outperforming a supervised Wide ResNet, and 99.0% accuracy with full finetuning, matching the top supervised pre-trained models. An even larger model trained on a mixture of ImageNet and web images is competitive with self-supervised benchmarks on ImageNet, achieving 72.0% top-1 accuracy on a linear probe of our features.","url_abs":"https://openai.com/blog/image-gpt/","url_pdf":"https://cdn.openai.com/papers/Generative_Pretraining_from_Pixels_V2.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":"generative-pretraining-from-pixels","repo_url":"https://github.com/openai/image-gpt","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"tf","reach":null},{"paper_slug":"generative-pretraining-from-pixels","repo_url":"https://github.com/EugenHotaj/pytorch-generative/blob/master/pytorch_generative/models/autoregressive/image_gpt.py","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null},{"paper_slug":"generative-pretraining-from-pixels","repo_url":"https://github.com/apeguero1/image-gpt/blob/master/Transformers_Image_GPT.ipynb","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"tf","reach":null},{"paper_slug":"generative-pretraining-from-pixels","repo_url":"https://github.com/teddykoker/image-gpt","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"representation-learning","task_name":"Representation Learning"},{"task_slug":"self-supervised-image-classification","task_name":"Self-Supervised Image Classification"}],"methods":[{"method_slug":"1x1-convolution","method_name":"1x1 Convolution"},{"method_slug":"adam","method_name":"Adam"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"average-pooling","method_name":"Average Pooling"},{"method_slug":"bpe","method_name":"BPE"},{"method_slug":"batch-normalization","method_name":"Batch Normalization"},{"method_slug":"bottleneck-residual-block","method_name":"Bottleneck Residual Block"},{"method_slug":"convolution","method_name":"Convolution"},{"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":"gpt-2","method_name":"GPT-2"},{"method_slug":"global-average-pooling","method_name":"Global Average Pooling"},{"method_slug":"kaiming-initialization","method_name":"Kaiming Initialization"},{"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":"max-pooling","method_name":"Max Pooling"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"residual-block","method_name":"Residual Block"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"weight-decay","method_name":"Weight Decay"},{"method_slug":"wide-residual-block","method_name":"Wide Residual Block"},{"method_slug":"wideresnet","method_name":"WideResNet"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/image-classification-on-stl-10","task":"Image Classification","dataset":"STL-10","model":"iGPT-L","rank_in_archive_order":14,"of":117,"metrics":{"Percentage correct":"97.1"},"uses_additional_data":true},{"leaderboard":"/sota/image-classification-on-stl-10","task":"Image Classification","dataset":"STL-10","model":"AMDIM-L","rank_in_archive_order":23,"of":117,"metrics":{"Percentage correct":"94.2"},"uses_additional_data":true},{"leaderboard":"/sota/self-supervised-image-classification-on","task":"Self-Supervised Image Classification","dataset":"ImageNet","model":"iGPT-XL (64x64, 15360 features)","rank_in_archive_order":96,"of":144,"metrics":{"Number of Params":"6801M","Top 1 Accuracy":"72.0%"},"uses_additional_data":false},{"leaderboard":"/sota/self-supervised-image-classification-on","task":"Self-Supervised Image Classification","dataset":"ImageNet","model":"iGPT-XL (64x64, 3072 features)","rank_in_archive_order":104,"of":144,"metrics":{"Number of Params":"6800M","Top 1 Accuracy":"68.7%"},"uses_additional_data":false},{"leaderboard":"/sota/self-supervised-image-classification-on","task":"Self-Supervised Image Classification","dataset":"ImageNet","model":"iGPT-L (48x48)","rank_in_archive_order":113,"of":144,"metrics":{"Number of Params":"1400M","Top 1 Accuracy":"65.2%"},"uses_additional_data":false},{"leaderboard":"/sota/self-supervised-image-classification-on","task":"Self-Supervised Image Classification","dataset":"ImageNet","model":"iGPT-L (32x32)","rank_in_archive_order":124,"of":144,"metrics":{"Number of Params":"1400M","Top 1 Accuracy":"60.3%"},"uses_additional_data":false}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}