# Less Is More: Linear Layers on CLIP Features as Powerful VizWiz Model Source: [Less Is More: Linear Layers on CLIP Features as Powerful VizWiz Model]() Collection index: [llms.txt]() Data updated: 2026-01-04T03:38:08.258Z Authors: Fabian Deuser, Konrad Habel, Philipp J. R\\"osch, Norbert Oswald Published: 2022-06-10T07:03:52.000Z ## Abstract Current architectures for multi-modality tasks such as visual question answering suffer from their high complexity. As a result, these architectures are difficult to train and require high computational resources. To address these problems we present a CLIP-based architecture that does not require any fine-tuning of the feature extractors. A simple linear classifier is used on the concatenated features of the image and text encoder. During training an auxiliary loss is added which operates on the answer types. The resulting classification is then used as an attention gate on the answer class selection. On the VizWiz 2022 Visual Question Answering Challenge we achieve 60.15 % accuracy on Task 1: Predict Answer to a Visual Question and AP score of 83.78 % on Task 2: Predict Answerability of a Visual Question. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Answerability on VizWiz 2022 (test-dev)](): Answerability; VizWiz 2022 (test-dev); AP: 84.13. - [Answerability on VizWiz 2022 (test)](): Answerability; VizWiz 2022 (test); AP: 83.78. - [Visual Question Answering on VizWiz 2022 (test-dev)](): Visual Question Answering; VizWiz 2022 (test-dev); Accuracy: 61.64. - [Visual Question Answering on VizWiz 2022 (test-std)](): Visual Question Answering; VizWiz 2022 (test-std); Accuracy: 60.15. ## Original sources and implementation [Original paper]() [PDF]()