CoolFace
Datasetpublic

xjtupanda/VisualProbe_Hard

VLMsAreBlind Benchmark refactor VisualProbe Benchmark to add support for VLMEvalKit. Benchmark Information Number of questions: 106 Question type: Free-form Question format: image + text Answer type: Free-form Reference VLMEvalKit VisualProbe Benchmark

sourceHugging Faceupdated 1y agoView on Hugging Face
0likes49downloads
Dataset Card

VLMsAreBlind Benchmark

refactor VisualProbe Benchmark to add support for VLMEvalKit.

Benchmark Information

  1. 1.Number of questions: 106
  2. 2.Question type: Free-form
  3. 3.Question format: image + text
  4. 4.Answer type: Free-form

Reference