paper-with-me

Papers

Generative Pretraining from Pixels

2020-07-17 · ICML 2020 1 · Mark Chen, Alec Radford, Rewon Child, Jeff Wu, Heewoo Jun, Prafulla Dhariwal, David Luan, Ilya Sutskever

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.

📄 PDF Abstract BibTeX

Code (4)

openai/image-gpt 공식 구현 tf
EugenHotaj/pytorch-generative/blob/master/pytorch_generative/models/autoregressive/image_gpt.py pytorch
apeguero1/image-gpt/blob/master/Transformers_Image_GPT.ipynb tf
teddykoker/image-gpt pytorch

Tasks

Image ClassificationRepresentation LearningSelf-Supervised Image Classification

Methods 이 논문이 사용한 방법론

Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
1x1 Convolution A 1 x 1 Convolution is a convolution with some special properties in that it can be used for dimensionality reduction,…
Cosine Annealing Cosine Annealing is a type of learning rate schedule that has the effect of starting with a large learning rate that is relatively rapidly decreased to a minimum value before…
Wide Residual Block 설명 없음
WideResNet 설명 없음
Weight Decay 설명 없음
Bottleneck Residual Block A Bottleneck Residual Block is a variant of the residual block that utilises 1x1 convolutions to create a bottleneck. The…
Batch Normalization 설명 없음

Similar Papers 제목 키워드 기반

End-to-end Generative Pretraining for Multimodal Video Captioning

2022-01-20 · CVPR 2022 1 · Paul Hongsuck Seo, Arsha Nagrani, Anurag Arnab, Cordelia Schmid

Recent video and language pretraining frameworks lack the ability to generate sentences. We present Multimodal Video Generative Pretraining (MV-GPT), a new pretraining framework for learning from unlabelled videos which …

Action ClassificationDecoderRetrievalSentence+3

Effective End-to-End Vision Language Pretraining with Semantic Visual Loss

2023-01-18 · Xiaofeng Yang, Fayao Liu, Guosheng Lin

Current vision language pretraining models are dominated by methods using region visual features extracted from object detectors. Given their good performance, the extract-then-process pipeline significantly restricts th…

GPU

CP2: Copy-Paste Contrastive Pretraining for Semantic Segmentation

2022-03-22 · Feng Wang, Huiyu Wang, Chen Wei, Alan Yuille 외

Recent advances in self-supervised contrastive learning yield good image-level representation, which favors classification tasks but usually neglects pixel-level detailed information, leading to unsatisfactory transfer p…

Contrastive LearningRepresentation LearningSegmentationSemantic Segmentation

Learning from Next-Frame Prediction: Autoregressive Video Modeling Encodes Effective Representations

2025-12-24 · Jinghan Li, Yang Jin, Hao Jiang, Yadong Mu 외 arxiv

Recent advances in pretraining general foundation models have significantly improved performance across diverse downstream tasks. While autoregressive (AR) generative models like GPT have revolutionized NLP, most visual …

Representation Learning

Towards Understanding Why Mask-Reconstruction Pretraining Helps in Downstream Tasks

2022-06-08 · Jiachun Pan, Pan Zhou, Shuicheng Yan

For unsupervised pretraining, mask-reconstruction pretraining (MRP) approaches, e.g. MAE and data2vec, randomly mask input patches and then reconstruct the pixels or semantic features of these masked patches via an auto-…

Decoder