REVIEW 4 cited by
How to Train Vision Transformer on Small-scale Datasets?
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Signed reviews
read the original abstract
Vision Transformer (ViT), a radically different architecture than convolutional neural networks offers multiple advantages including design simplicity, robustness and state-of-the-art performance on many vision tasks. However, in contrast to convolutional neural networks, Vision Transformer lacks inherent inductive biases. Therefore, successful training of such models is mainly attributed to pre-training on large-scale datasets such as ImageNet with 1.2M or JFT with 300M images. This hinders the direct adaption of Vision Transformer for small-scale datasets. In this work, we show that self-supervised inductive biases can be learned directly from small-scale datasets and serve as an effective weight initialization scheme for fine-tuning. This allows to train these models without large-scale pre-training, changes to model architecture or loss functions. We present thorough experiments to successfully train monolithic and non-monolithic Vision Transformers on five small datasets including CIFAR10/100, CINIC10, SVHN, Tiny-ImageNet and two fine-grained datasets: Aircraft and Cars. Our approach consistently improves the performance of Vision Transformers while retaining their properties such as attention to salient regions and higher robustness. Our codes and pre-trained models are available at: https://github.com/hananshafi/vits-for-small-scale-datasets.
Forward citations
Cited by 4 Pith papers
-
Benign Overfitting Does Not Occur in Diffusion Models
Benign overfitting and double descent do not occur in diffusion models: population and empirical score-matching losses cannot both be small without exponentially many samples.
-
Cyborg Insect Factory: Automatic Assembly System to Build up Insect-computer Hybrid Robot Based on Vision-guided Robotic Arm Manipulation of Custom Bipolar Electrodes
An automatic vision-guided robotic assembly system builds steerable and decelerable cyborg cockroaches in 68 seconds with control performance comparable to manual assembly.
-
Learning to Adapt to Position Bias in Vision Transformer Classifiers
A vision transformer can learn a single scalar that gates its position embedding, and a new SHAP-based metric measures when that gate should be small or large.
-
Powerful Design of Small Vision Transformer on CIFAR10
A Tiny ViT on CIFAR-10 reaches 93.95% accuracy with two classifier tokens at reduced width, while low-rank query compression causes only a small accuracy drop.
Discussion (0). Continue with ORCID to comment.