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

Recent years have witnessed remarkable progress in the development of largevision-language models (LVLMs). Benefiting from the strong language backbonesand efficient cross-modal alignment strategies, LVLMs exhibit surprisingcapabilities to perceive visual signals and perform visually groundedreasoning. However, the capabilities of LVLMs have not been comprehensively andquantitatively evaluate. Most existing multi-modal benchmarks requiretask-oriented input-output formats, posing great challenges to automaticallyassess the free-form text output of LVLMs. To effectively leverage theannotations available in existing benchmarks and reduce the manual effortrequired for constructing new benchmarks, we propose to re-formulate existingbenchmarks into unified LVLM-compatible formats. Through systematic datacollection and reformulation, we present the ReForm-Eval benchmark, offeringsubstantial data for evaluating various capabilities of LVLMs. Based onReForm-Eval, we conduct extensive experiments, thoroughly analyze the strengthsand weaknesses of existing LVLMs, and identify the underlying factors. Ourbenchmark and evaluation framework will be open-sourced as a cornerstone foradvancing the development of LVLMs.

ReForm-Eval: Evaluating Large Vision Language Models via Unified Re-Formulation of Task-Oriented Benchmarks