Papers with Code · CC BY-SA 4.0

Something-Something V2

아카이브 평가 123건 · 마지막 적재/변경 2026-10-30

평가 기록

과거 수집 기록이며 최신 순위가 아닙니다. 데이터셋 버전·split·실험 조건과 지표 방향이 검증되지 않은 결과는 직접 비교할 수 없습니다. 날짜는 논문 날짜이며 최신 평가일을 뜻하지 않습니다.

표가 화면보다 넓으면 좌우로 스크롤하여 논문과 구현 링크를 확인하세요.

원본 기록 순서 · 순위 미부여
모델	원본 지표·값	논문	구현
MVD (Kinetics400 pretrain, ViT-H, 16 frame)	Top-1 Accuracy 77.3 Top-5 Accuracy 95.7 Parameters 633 GFLOPs 1192x6	Masked Video Distillation: Rethinking Masked Feature Modeling for Self-supervised Video Representation Learning 2022-12-08	코드 1 · 코드 2 · 코드 3 · 코드 4
DejaVid	Top-1 Accuracy 77.2 Top-5 Accuracy 96.3	DejaVid: Encoder-Agnostic Learned Temporal Matching for Video Classification 2025-01-01	코드 1
InternVideo	Top-1 Accuracy 77.2	InternVideo: General Video Foundation Models via Generative and Discriminative Learning 2022-12-06	코드 1 · 코드 2
InternVideo2-1B	Top-1 Accuracy 77.1	InternVideo2: Scaling Foundation Models for Multimodal Video Understanding 2024-03-22	코드 1 · 코드 2
VideoMAE V2-g	Top-1 Accuracy 77.0 Top-5 Accuracy 95.9 Parameters 1013 GFLOPs 2544x6	VideoMAE V2: Scaling Video Masked Autoencoders with Dual Masking 2023-03-29	코드 1
MVD (Kinetics400 pretrain, ViT-L, 16 frame)	Top-1 Accuracy 76.7 Top-5 Accuracy 95.5 Parameters 305 GFLOPs 597x6	Masked Video Distillation: Rethinking Masked Feature Modeling for Self-supervised Video Representation Learning 2022-12-08	코드 1 · 코드 2 · 코드 3 · 코드 4
Hiera-L (no extra data)	Top-1 Accuracy 76.5	Hiera: A Hierarchical Vision Transformer without the Bells-and-Whistles 2023-06-01	코드 1 · 코드 2 · 코드 3 · 코드 4
TubeViT-L	Top-1 Accuracy 76.1 Top-5 Accuracy 95.2	Rethinking Video ViTs: Sparse Video Tubes for Joint Image and Video Learning 2022-12-06	코드 1
VideoMAE (no extra data, ViT-L, 32x2)	Top-1 Accuracy 75.4 Top-5 Accuracy 95.2 Parameters 305 GFLOPs 1436x3	VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training 2022-03-23	코드 1 · 코드 2 · 코드 3 · 코드 4 · 코드 5 · 코드 6 · 코드 7 · 코드 8 · 코드 9
Side4Video (EVA ViT-E/14)	Top-1 Accuracy 75.2 Top-5 Accuracy 94.0	Side4Video: Spatial-Temporal Side Network for Memory-Efficient Image-to-Video Transfer Learning 2023-11-27	코드 1 · 코드 2
MaskFeat (Kinetics600 pretrain, MViT-L)	Top-1 Accuracy 75.0 Top-5 Accuracy 95.0 Parameters 218 GFLOPs 2828*3	Masked Feature Prediction for Self-Supervised Visual Pre-Training 2021-12-16	코드 1 · 코드 2 · 코드 3 · 코드 4 · 코드 5 · 코드 6
MAR (50% mask, ViT-L, 16x4)	Top-1 Accuracy 74.7 Top-5 Accuracy 94.9 Parameters 311 GFLOPs 276x6	MAR: Masked Autoencoders for Efficient Action Recognition 2022-07-24	코드 1
ATM	Top-1 Accuracy 74.6 Top-5 Accuracy 94.4	What Can Simple Arithmetic Operations Do for Temporal Modeling? 2023-07-18	코드 1 · 코드 2
MAWS (ViT-L)	Top-1 Accuracy 74.4	The effectiveness of MAE pre-pretraining for billion-scale pretraining 2023-03-23	코드 1
VideoMAE (no extra data, ViT-L, 16frame)	Top-1 Accuracy 74.3 Top-5 Accuracy 94.6 Parameters 305 GFLOPs 597x6	VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training 2022-03-23	코드 1 · 코드 2 · 코드 3 · 코드 4 · 코드 5 · 코드 6 · 코드 7 · 코드 8 · 코드 9
MAR (75% mask, ViT-L, 16x4)	Top-1 Accuracy 73.8 Top-5 Accuracy 94.4 Parameters 311 GFLOPs 131x6	MAR: Masked Autoencoders for Efficient Action Recognition 2022-07-24	코드 1
MVD (Kinetics400 pretrain, ViT-B, 16 frame)	Top-1 Accuracy 73.7 Top-5 Accuracy 94.0 Parameters 87 GFLOPs 180x6	Masked Video Distillation: Rethinking Masked Feature Modeling for Self-supervised Video Representation Learning 2022-12-08	코드 1 · 코드 2 · 코드 3 · 코드 4
ViC-MAE (ViT-L)	Top-1 Accuracy 73.7	ViC-MAE: Self-Supervised Representation Learning from Images and Video with Contrastive Masked Autoencoders 2023-03-21	코드 1 · 코드 2
TAdaFormer-L/14	Top-1 Accuracy 73.6	Temporally-Adaptive Models for Efficient Video Understanding 2023-08-10	코드 1
TDS-CLIP-ViT-L/14(8frames)	Top-1 Accuracy 73.4 Top-5 Accuracy 93.8	TDS-CLIP: Temporal Difference Side Network for Image-to-Video Transfer Learning 2024-08-20	코드 1
MViTv2-L (IN-21K + Kinetics400 pretrain)	Top-1 Accuracy 73.3 Top-5 Accuracy 94.1 Parameters 213.1	MViTv2: Improved Multiscale Vision Transformers for Classification and Detection 2021-12-02	코드 1 · 코드 2 · 코드 3 · 코드 4 · 코드 5 · 코드 6 · 코드 7 · 코드 8 · 코드 9
AMD(ViT-B/16)	Top-1 Accuracy 73.3 Top-5 Accuracy 94.0 Parameters 87 GFLOPs 180x6	Asymmetric Masked Distillation for Pre-Training Small Foundation Models 2023-11-06	연결 없음
UniFormerV2-L	Top-1 Accuracy 73.0 Top-5 Accuracy 94.5 GFLOPs 5154	UniFormerV2: Spatiotemporal Learning by Arming Image ViTs with Video UniFormer 2022-09-22	코드 1 · 코드 2
ST-Adapter (ViT-L, CLIP)	Top-1 Accuracy 72.3 Top-5 Accuracy 93.9 GFLOPs 8248	ST-Adapter: Parameter-Efficient Image-to-Video Transfer Learning 2022-06-27	코드 1
ZeroI2V ViT-L/14	Top-1 Accuracy 72.2 Top-5 Accuracy 93.0	ZeroI2V: Zero-Cost Adaptation of Pre-trained Transformers from Image to Video 2023-10-02	코드 1 · 코드 2
MViT-B (IN-21K + Kinetics400 pretrain)	Top-1 Accuracy 72.1 GFLOPs 225x3	MViTv2: Improved Multiscale Vision Transformers for Classification and Detection 2021-12-02	코드 1 · 코드 2 · 코드 3 · 코드 4 · 코드 5 · 코드 6 · 코드 7 · 코드 8 · 코드 9
CAST(ViT-B/16)	Top-1 Accuracy 71.6	CAST: Cross-Attention in Space and Time for Video Action Recognition 2023-11-30	코드 1
StructVit-B-4-1	Top-1 Accuracy 71.5	Learning Correlation Structures for Vision Transformers 2024-04-05	연결 없음
OMNIVORE (Swin-B, IN-21K+ Kinetics400 pretrain)	Top-1 Accuracy 71.4 Top-5 Accuracy 93.5	Omnivore: A Single Model for Many Visual Modalities 2022-01-20	코드 1 · 코드 2
BEVT (IN-1K + Kinetics400 pretrain)	Top-1 Accuracy 71.4 Top-5 Accuracy - Parameters 89 GFLOPs 321x3	BEVT: BERT Pretraining of Video Transformers 2021-12-02	코드 1