List your company
ResearchFormerly WizWand

Open-ended Video Question Answering on MSVD-QA

79.9AccuracyPLLaVA
-3.09218.4544061.546Jun 16, 2022Nov 20, 2022Apr 26, 2023Sep 30, 2023Mar 5, 2024Aug 9, 2024Jan 14, 2025
Updated 8mo ago

Evaluation results

MethodLinks
PLLaVALLM Size=34B, Vision Encoder=CLIP-L2024.0779.94.2
SF-LLaVA-34BLLM Size=34B, Vision Encoder=CLIP-L2024.0779.94.1
PLLAVALLM Size=34B, Vision Encoder=CLIP-L, Training-free=false2024.1179.94.2
SF-LLAVALLM Size=34B, Vision Encoder=CLIP-L, Training-free=true2024.1179.94.1
IG-VLM (LLaVA-v1.6)LLM Size=34B, Vision Encoder=CLIP-L2024.0779.64.1
IG-VLMLLM Size=34B, Vision Encoder=CLIP-L, Training-free=true2024.1179.64.1
TS-LLAVALLM Size=34B, Vision Encoder=CLIP-L, Training-free=true2024.1179.44.1
SF-LLaVALLM Size=7B, Vision Encoder=CLIP-L, Training-free=true2024.1179.14.1
TS-LLaVALLM Size=7B, Vision Encoder=CLIP-L, Training-free=true2024.11794.1
IG-VLMLLM Size=7B, Vision Encoder=CLIP-L, Training-free=true2024.1178.84.1
PLLAVALLM Size=7B, Vision Encoder=CLIP-L, Training-free=false2024.1176.64.1
P-LLaVAParameters=7B2025.0176.64.1
MovieChat+LLM Size=7B, Vision Encoder=CLIP-G, Training-free=false2024.1176.53.9
Grounded-VideoLLM2025.0176.34.1
DeepStack-LLLM Size=7B, Vision Encoder=CLIP-L, Training-free=true2024.11764
LLaVA-ST2025.0175.94.1
Elysium2025.0175.83.7
MovieChatLLM Size=7B, Vision Encoder=CLIP-G, Training-free=false2024.1175.23.8
MovieChat2025.0175.23.8
ST-LLM2025.0174.63.9
FreeVALLM Size=7B, Vision Encoder=CLIP-L, Training-free=true2024.1173.84.1
VideoGPT+LLM Size=3.8B, Vision Encoder=CLIP-L, Training-free=false2024.1172.43.9
Baichuan-omni (7B)# Frames=1 fps (max 48)2024.1072.24
Video-LLaMA2LLM Size=7B, Vision Encoder=CLIP-L, Training-free=false2024.1170.93.8
VideoLLaMA 2 (7B)# Frames=162024.1070.93.8
Video-LLaVALLM Size=7B, Vision Encoder=ViT-L, Training-free=false2024.1170.73.9
Video-LLaVA (7B)# Frames=82024.1070.73.9
Video-LLaMA2LLM Size=46.7B, Vision Encoder=CLIP-L2024.0770.53.8
Video-LLaMA2LLM Size=46.7B, Vision Encoder=CLIP-L, Training-free=false2024.1170.53.8
VideoChat2LLM Size=7B, Vision Encoder=UMT-L, Training-free=false2024.11703.9
VideoChat2 (7B)# Frames=162024.10703.9
LongVLM2025.01703.8
VideoChat22025.01703.9
LLaMA-VIDLLM Size=13B, Vision Encoder=CLIP-G, Training-free=false2024.1169.73.7
Momentor2025.0168.93.6
GroundingGPT2025.0167.83.7
LLaVA-NeXT-Video (7B)# Frames=322024.1067.43.4
Vista-LLaMALLM Size=7B, Vision Encoder=CLIP-G, Training-free=false2024.1165.33.6
Chat-UniVi2025.01653.6
Video-ChatGPTLLM Size=7B, Vision Encoder=CLIP-L, Training-free=false2024.1164.93.3
Video-ChatGPT2025.0164.93.3
VITA (8x7B)# Frames=1 fps (max 32)2024.1063.93.7
Qwen2 VL (7B)# Frames=2 fps (max 768)2024.1061.13.5
VideoChatLLM Size=7B, Vision Encoder=CLIP-G, Training-free=false2024.1156.32.8
FrozenBiLMSupervision=100% (fully-supervised)2022.0654.8—
Video-LLaMALLM Size=7B, Vision Encoder=CLIP-G, Training-free=false2024.1151.62.5
Video-LLaMA2025.0151.62.5
FrozenBiLMSupervision=10% (few-shot)2022.0651—
FrozenBiLMSupervision=1% (few-shot)2022.0646.5—
FrozenBiLMTraining Data=WebVid10M, Speech=true, Zero-shot=true2022.0633.8—
FrozenBiLMSupervision=0% (zero-shot)2022.0633.8—
FrozenBiLMTraining Data=WebVid10M, Speech=false, Zero-shot=true2022.0633.7—
FrozenBiLMTraining Data=WebVid10M, Zero-shot evaluation=true2022.0630—
Just AskTraining Data=HowToVQA69M + WebVidVQA3M, Speech=false, Zero-shot=true2022.0613.5—
Just AskTraining Data=HowToVQA69M + WebVidVQA3M, Zero-shot evaluation=true2022.0613.5—
CLIP ViT-L/14Training Data=400M image-texts, Speech=false, Zero-shot=true2022.067.2—
CLIP ViT-L/14Training Data=400M image-texts, Zero-shot evaluation=true2022.067.2—
RandomZero-shot=true2022.060.1—
RandomTraining Data=—, Zero-shot evaluation=true2022.060.1—