List your company
ResearchFormerly WizWand

Speech-Visual Question Answering on VQA Speech-converted v2

32.95AccuracySingle-Modality-Expert-Task3
1.10529.372617.6425.9074May 18, 2024
Updated 8mo ago

Evaluation results

MethodLinks
Single-Modality-Expert-Task3Evaluation Protocol=zero-shot2024.0532.95
Single-Modality-Expert-Task1Evaluation Protocol=zero-shot2024.0532.6
Uni-MoETraining Task=MoE-Task1-short-speech, Training Epochs=2 epoch2024.0531.28
Uni-MoETraining Task=MoE-Task1-short-speech, Training Epochs=3 epoch2024.0530.86
X-InstructBLIPEvaluation Protocol=zero-shot2024.0526.47
Uni-MoETraining Task=MoE-Task1, Training Epochs=2 epoch2024.0526.01
Single-Modality-Expert-Task6Evaluation Protocol=zero-shot2024.0525.11
Single-Modality-Expert-Task4Evaluation Protocol=zero-shot2024.0524.3
Uni-MoETraining Task=MoE-Task1, Training Epochs=3 epoch2024.0520.14
Uni-MoETraining Task=MoE-Task2, Training Epochs=2 epoch2024.0512.68
Single-Modality-Expert-Task5Evaluation Protocol=zero-shot2024.055.9
Macaw-LLMEvaluation Protocol=zero-shot2024.052.33