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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Large Language Models (LLMs), primarily trained on text-based datasets,exhibit exceptional proficiencies in understanding and executing complexlinguistic instructions via text outputs. However, they falter when requests togenerate non-text ones. Concurrently, modality conversion models, such astext-to-image, despite generating high-quality images, suffer from a lack ofextensive textual pretraining. As a result, these models are only capable ofaccommodating specific image descriptions rather than comprehending morecomplex instructions. To bridge this gap, we propose a novel approach,\methodname, from a modality conversion perspective that evolves a text-basedLLM into a multi-modal one. We specifically employ a minimal dataset toinstruct LLMs to recognize the intended output modality as directed by theinstructions. Consequently, the adapted LLM can effectively summon variousoff-the-shelf modality conversion models from the model zoos to generatenon-text responses. This circumvents the necessity for complicated pretrainingthat typically requires immense quantities of paired multi-modal data, whilesimultaneously inheriting the extensive knowledge of LLMs and the ability ofhigh-quality generative models. To evaluate and compare the adapted multi-modalLLM with its traditional counterparts, we have constructed a multi-modalinstruction benchmark that solicits diverse modality outputs. The experimentresults reveal that, with minimal training, LLMs can be conveniently adapted tocomprehend requests for non-text responses, thus achieving higher flexibilityin multi-modal scenarios. Code and data will be made available athttps://github.com/xinke-wang/SwitchGPT.

SwitchGPT: Adapting Large Language Models for Non-Text Outputs