English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In recent developments in the field of Computer Vision, a rise is seen in theuse of transformer-based architectures. They are surpassing thestate-of-the-art set by CNN architectures in accuracy but on the other hand,they are computationally very expensive to train from scratch. As these modelsare quite recent in the Computer Vision field, there is a need to study it'stransfer learning capabilities and compare it with CNNs so that we canunderstand which architecture is better when applied to real world problemswith small data. In this work, we follow a simple yet restrictive method forfine-tuning both CNN and Transformer models pretrained on ImageNet1K onCIFAR-10 and compare them with each other. We only unfreeze the lasttransformer/encoder or last convolutional block of a model and freeze all thelayers before it while adding a simple MLP at the end for classification. Thissimple modification lets us use the raw learned weights of both these neuralnetworks. From our experiments, we find out that transformers-basedarchitectures not only achieve higher accuracy than CNNs but some transformerseven achieve this feat with around 4 times lesser number of parameters.

Investigating Transfer Learning Capabilities of Vision Transformers and CNNs by Fine-Tuning a Single Trainable Block