List your company

Video Question Answering on How2QA

93.2AccT+T (+ASR)
39.84853.69967.5581.401Oct 12, 2022Jan 2, 2023Mar 25, 2023Jun 15, 2023Sep 5, 2023Nov 26, 2023Feb 16, 2024

Evaluation results

MethodLinks
T+T (+ASR)Input=dense/1fps, Speech Input=true2023.0593.2
T+TInput=dense/1fps, Speech Input=true2023.0592.4
FrozenBiLMNumber of Frames=10, Speech Input=true2023.0586.7
JustAskNumber of Frames=20, Speech Input=true2023.0584.4
JustAskzero-shot=false2023.1284.4
ViLAFrames Numbers=42023.1283.9
SEVILA†Number of Frames=32 -> 4, Keyframe Selection=zero-shot LOCALIZER2023.0583.7
SEVILANumber of Frames=32 -> 4, Keyframe Selection=refined LOCALIZER2023.0583.6
SeViLAFrames Numbers=42023.1283.6
SeViLAzero-shot=false2023.1283.6
ViLAFrames Numbers=22023.1282.8
BLIP-2 concat (ANSWERER)Number of Frames=4, Protocol=feature concatenation2023.0582.2
BLIP-2Frames Numbers=42023.1282.2
FrozenBiLMNumber of Frames=102023.0581.5
FrozenBiLMFrames Numbers=102023.1281.5
BLIP-2 votingNumber of Frames=4, Protocol=majority voting2023.0579.6
InternVideoNumber of Frames=8, Tested_by_authors=true2023.0579
InternVideoFrames Numbers=82023.1279
LF-VILA2022.1076.1
ResNet-SF2022.1074.3
HERO2022.1074.3
HEROInput=dense/1fps, Speech Input=true2023.0573.8
CLIP-SF2022.1072.9
SeViLA†Number of Frames=32 -> 4, Zero-shot=true2023.0572.3
SeViLaevaluation_mode=Zero-shot2024.0272.3
Q-ViDevaluation_mode=Zero-shot2024.0271.4
BLIP-2concat (Answerer)Number of Frames=4, Zero-shot=true2023.0570.8
BLIP-2concatevaluation_mode=Zero-shot2024.0270.8
BLIPv2 ViTG FlanT5xxlzero-shot=true, vision encoder=ViTG, LLM=FlanT5xxl, number of frames=42023.0969.8
BLIP-2votingNumber of Frames=4, Zero-shot=true2023.0569.8
BLIP-2votingevaluation_mode=Zero-shot2024.0269.8
CLIPimplementation_source=[28]2022.1069.3
AnyMAL-Image 70Bzero-shot=true, vision encoder=ViT-G, LLM size=70B, number of frames=42023.0968.1
OneLLM-7Bzero-shot=true2023.1265.7
Internvideozero-shot=true, number of frames=82023.0962.2
InternVideo*Number of Frames=8, Zero-shot=true2023.0562.2
InternVideoevaluation_mode=Zero-shot2024.0262.2
InternVideozero-shot=true2023.1262.2
AnyMAL-Video 70Bzero-shot=true, vision encoder=Internvideo, LLM size=70B, number of frames=82023.0960
AnyMAL-Image 13Bzero-shot=true, vision encoder=ViT-G, LLM size=13B, number of frames=42023.0959.6
AnyMAL-13Bzero-shot=true2023.1259.6
FrozenBiLMNumber of Frames=10, speech input or dense frames=true, Zero-shot=true2023.0558.4
FrozenBiLMzero-shot=true2023.1258.4
AnyMAL-Video 13Bzero-shot=true, vision encoder=Internvideo, LLM size=13B, number of frames=82023.0954.8
JustAskNumber of Frames=20, speech input or dense frames=true, Zero-shot=true2023.0551.1
FrozenBiLMNumber of Frames=10, Zero-shot=true2023.0541.9
FrozenBILMevaluation_mode=Zero-shot2024.0241.9