List your company

Movie Audio Description generation on MAD-eval-Named v2 (test)

28.2C ScoreUni-AD
3.4489.87416.322.726Mar 29, 2023May 27, 2023Jul 25, 2023Sep 23, 2023Nov 21, 2023Jan 19, 2024Mar 19, 2024

Evaluation results

MethodLinks
Uni-ADContextual Information=true, Language Model=LLaMA2, Visual Feature=CLIP-L142024.0328.217.1——54.2
Uni-ADContextual Information=false, Language Model=LLaMA2, Visual Feature=CLIP-L142024.0327.316.8——53.3
Uni-ADContextual Information=false, Language Model=GPT-2, Visual Feature=CLIP-L142024.0325.716.4——51.5
AutoAD-IIIContextual Information=false, Language Model=LLaMA2, Visual Feature=EVA-CLIP2024.0324———52.8
Uni-ADContextual Information=false, Language Model=GPT-2, Visual Feature=CLIP-B322024.032415.9——50.5
AutoAD-ZeroContextual Information=false, Language Model=LLaMA3, Visual Feature=VideoLLaMA22024.0322.4————
AutoAD-IIContextual Information=true, Language Model=GPT-2, Visual Feature=CLIP-B322024.0319.513.4——50.8
AutoAD-IIContextual Information=false, Language Model=GPT-2, Visual Feature=CLIP-B322024.0319.213.1——51.3
AutoADPretraining Data=AV-AD & WebVid2023.0314.311.94.424.2—
AutoAD-IContextual Information=true, Language Model=GPT-2, Visual Feature=CLIP-B322024.0314.311.9——42.1
AutoADPretraining Data=AV-AD2023.0314.112.14.223—
MM-NarratorContextual Information=true, Language Model=GPT-4, Visual Feature=CLIP-L142024.0313.915.1——49
AutoAD-IContextual Information=false, Language Model=GPT-2, Visual Feature=CLIP-B322024.0312.110.3——39.8
MM-NarratorContextual Information=true, Language Model=GPT-4V, Visual Feature=CLIP-L142024.039.812.8———
CapDec*Pretraining Data=AV-AD2023.036.78.21.414.3—
ClipCapPretraining Data=CC3M2023.034.48.51.111.8—
ClipCapContextual Information=false, Language Model=GPT-2, Visual Feature=CLIP-B322024.034.48.5——36.5