Video Based Generative Performance

Métriques

Consistency

Contextual Understanding

Correctness of Information

Detail Orientation

Temporal Understanding

mean

Résultats

Résultats de performance de divers modèles sur ce benchmark

Nom du modèle	Consistency	Contextual Understanding	Correctness of Information	Detail Orientation	Temporal Understanding	mean	Paper Title	Repository
VideoChat2_HD_mistral	2.84	3.72	3.40	2.91	2.65	3.10	MVBench: A Comprehensive Multi-modal Video Understanding Benchmark	-
BT-Adapter (zero-shot)	2.2	2.89	2.16	2.46	2.13	2.46	BT-Adapter: Video Conversation is Feasible Without Video Instruction Tuning	-
TS-LLaVA-34B	-	-	-	-	-	3.38	TS-LLaVA: Constructing Visual Tokens through Thumbnail-and-Sampling for Training-Free Video Large Language Models	-
LLaMA-VID-7B (2 Token)	2.51	3.53	2.96	3.00	2.46	2.89	LLaMA-VID: An Image is Worth 2 Tokens in Large Language Models	-
LLaMA-VID-13B (2 Token)	2.63	3.60	3.07	3.05	2.58	2.99	LLaMA-VID: An Image is Worth 2 Tokens in Large Language Models	-
LLaMA Adapter	2.15	2.30	2.03	2.32	1.98	2.16	LLaMA-Adapter V2: Parameter-Efficient Visual Instruction Model	-
BT-Adapter	2.46	3.27	2.68	2.69	2.34	2.69	BT-Adapter: Video Conversation is Feasible Without Video Instruction Tuning	-
VLM-RLAIF	3.32	4	3.63	3.25	3.23	3.49	Tuning Large Multimodal Models for Videos using Reinforcement Learning from AI Feedback	-
VideoChat2	2.81	3.51	3.02	2.88	2.66	2.98	MVBench: A Comprehensive Multi-modal Video Understanding Benchmark	-
CAT-7B	2.89	3.49	3.08	2.95	2.81	3.07	CAT: Enhancing Multimodal Large Language Model to Answer Questions in Dynamic Audio-Visual Scenarios	-
PPLLaVA-7B	3.20	3.88	3.32	3.20	3.0	3.32	PPLLaVA: Varied Video Sequence Understanding With Prompt Guidance	-
PLLaVA-34B	3.25	3.90	3.60	3.20	2.67	3.32	PLLaVA : Parameter-free LLaVA Extension from Images to Videos for Video Dense Captioning	-
VideoGPT+	3.39	3.74	3.27	3.18	2.83	3.28	VideoGPT+: Integrating Image and Video Encoders for Enhanced Video Understanding	-
Chat-UniVi	2.81	3.46	2.89	2.91	2.39	2.99	Chat-UniVi: Unified Visual Representation Empowers Large Language Models with Image and Video Understanding	-
LITA-13B	3.19	3.43	2.94	2.98	2.68	3.04	LITA: Language Instructed Temporal-Localization Assistant	-
PPLLaVA-7B-dpo	3.81	4.21	3.85	3.56	3.21	3.73	PPLLaVA: Varied Video Sequence Understanding With Prompt Guidance	-
Video Chat	2.24	2.53	2.23	2.50	1.94	2.29	VideoChat: Chat-Centric Video Understanding	-
SlowFast-LLaVA-34B	-	-	-	-	-	3.32	SlowFast-LLaVA: A Strong Training-Free Baseline for Video Large Language Models	-
ST-LLM-7B	2.81	3.74	3.23	3.05	2.93	3.15	ST-LLM: Large Language Models Are Effective Temporal Learners	-
Video LLaMA	1.79	2.16	1.96	2.18	1.82	1.98	Video-LLaMA: An Instruction-tuned Audio-Visual Language Model for Video Understanding	-

0 of 23 row(s) selected.