Visual Question Answering (VQA) on ScanQA Test w/ objects

Metric: Exact Match (higher is better)

LeaderboardDataset

Loading chart...

Results

Sort:

#	Model↕	Exact Match▼	Extra Data	Paper	Date↕	Code
1	BridgeQA	31.29	No	Bridging the Gap between 2D and 3D Visual Questi...	2024-02-24	Code
2	LLaVA-3D	30.6	No	LLaVA-3D: A Simple yet Effective Pathway to Empo...	2024-09-26	-
3	Video-3D LLM	30.1	No	Video-3D LLM: Learning Position-Aware Video Repr...	2024-11-30	Code
4	Scene-LLM	27.2	No	Scene-LLM: Extending Language Model for 3D Visua...	2024-03-18	-
5	NaviLLM	26.27	No	Towards Learning a Generalist Model for Embodied...	2023-12-04	Code
6	LEO	24.5	No	An Embodied Generalist Agent in 3D World	2023-11-18	Code
7	ScanQA	23.45	No	ScanQA: 3D Question Answering for Spatial Scene ...	2021-12-20	Code
8	3D-LLM (flamingo)	23.2	No	3D-LLM: Injecting the 3D World into Large Langua...	2023-07-24	Code
9	3D-VisTA	22.4	No	3D-VisTA: Pre-trained Transformer for 3D Vision ...	2023-08-08	Code
10	ChatScene	21.6	No	Chat-Scene: Bridging 3D Scene and Large Language...	2023-12-13	Code
11	ScanRefer+MCAN	20.56	No	ScanQA: 3D Question Answering for Spatial Scene ...	2021-12-20	Code
12	VoteNet+MCAN	19.71	No	ScanQA: 3D Question Answering for Spatial Scene ...	2021-12-20	Code
13	VideoChat2	19.2	No	MVBench: A Comprehensive Multi-modal Video Under...	2023-11-28	Code
14	3D-LLM (BLIP2-flant5)	19.1	No	3D-LLM: Injecting the 3D World into Large Langua...	2023-07-24	Code
15	3D-LLM (BLIP2-opt)	19.1	No	3D-LLM: Injecting the 3D World into Large Langua...	2023-07-24	Code
16	LLaVA-NeXT-Video	18.7	No	LLaVA-OneVision: Easy Visual Task Transfer	2024-08-06	Code

#1BridgeQASOTA
31.29
Exact Match· 2024-02-24
Bridging the Gap between 2D and 3D Visual Question Answering: A Fusion Approach for 3D VQA Code
#2LLaVA-3D
30.6
Exact Match· 2024-09-26
LLaVA-3D: A Simple yet Effective Pathway to Empowering LMMs with 3D-awareness
#3Video-3D LLM
30.1
Exact Match· 2024-11-30
Video-3D LLM: Learning Position-Aware Video Representation for 3D Scene Understanding Code
#4Scene-LLM
27.2
Exact Match· 2024-03-18
Scene-LLM: Extending Language Model for 3D Visual Understanding and Reasoning
#5NaviLLMSOTA
26.27
Exact Match· 2023-12-04
Towards Learning a Generalist Model for Embodied Navigation Code
#6LEOSOTA
24.5
Exact Match· 2023-11-18
An Embodied Generalist Agent in 3D World Code
#7ScanQASOTA
23.45
Exact Match· 2021-12-20
ScanQA: 3D Question Answering for Spatial Scene Understanding Code
#83D-LLM (flamingo)
23.2
Exact Match· 2023-07-24
3D-LLM: Injecting the 3D World into Large Language Models Code
#93D-VisTA
22.4
Exact Match· 2023-08-08
3D-VisTA: Pre-trained Transformer for 3D Vision and Text Alignment Code
#10ChatScene
21.6
Exact Match· 2023-12-13
Chat-Scene: Bridging 3D Scene and Large Language Models with Object Identifiers Code
#11ScanRefer+MCAN
20.56
Exact Match· 2021-12-20
ScanQA: 3D Question Answering for Spatial Scene Understanding Code
#12VoteNet+MCAN
19.71
Exact Match· 2021-12-20
ScanQA: 3D Question Answering for Spatial Scene Understanding Code
#13VideoChat2
19.2
Exact Match· 2023-11-28
MVBench: A Comprehensive Multi-modal Video Understanding Benchmark Code
#143D-LLM (BLIP2-flant5)
19.1
Exact Match· 2023-07-24
3D-LLM: Injecting the 3D World into Large Language Models Code
#153D-LLM (BLIP2-opt)
19.1
Exact Match· 2023-07-24
3D-LLM: Injecting the 3D World into Large Language Models Code
#16LLaVA-NeXT-Video
18.7
Exact Match· 2024-08-06
LLaVA-OneVision: Easy Visual Task Transfer Code