# Action Recognition on Kinetics-400 (test)

Source: [Action Recognition on Kinetics-400 (test)](<https://www.sota2.com/research/sota/action-recognition-on-kinetics-400-test>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-03-20T09:36:03.393Z

Evaluation of video action recognition performance using the Kinetics-400 dataset test split, typically measured by Top-1 and Top-5 accuracy.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Action Recognition |
| Dataset (including any stated split) | Kinetics 400 (test) |
| Metric | Top-1 Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 270 of 270 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Top-1 Accuracy](<https://www.sota2.com/research/sota/action-recognition-on-kinetics-400-test.md?metric=Top-1%20Accuracy>): Higher is better; unit: PERCENT.

- [Top-5 Accuracy](<https://www.sota2.com/research/sota/action-recognition-on-kinetics-400-test.md?metric=Top-5%20Accuracy>): Higher is better; unit: PERCENT.

- [Top-10 Accuracy](<https://www.sota2.com/research/sota/action-recognition-on-kinetics-400-test.md?metric=Top-10%20Accuracy>): Higher is better; unit: PERCENT.

- [Accuracy](<https://www.sota2.com/research/sota/action-recognition-on-kinetics-400-test.md?metric=Accuracy>): Higher is better; unit: PERCENT.

- [EDM](<https://www.sota2.com/research/sota/action-recognition-on-kinetics-400-test.md?metric=EDM>): Lower is better; unit: SCALAR.

## Recorded results

| Method | Top-1 Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| UniFormerV2-L | 89.7 | Not recorded | 68fa75861e209131cea685ce |
| AM/12, Transformer | 89.6 | Not recorded | 68fa75861e209131cea685ce |
| MTV-H | 89.1 | Not recorded | 68fa75861e209131cea685ce |
| AM/12, Transformer | 89.1 | Not recorded | 68fa75861e209131cea685ce |
| CoCa | 88.9 | Not recorded | 68fa75861e209131cea685ce |
| InternVideo2s2-6B | 88.8 | Not recorded | 68fa74871e209131cea623f3 |
| AM/12, LSTM | 88.8 | Not recorded | 68fa75861e209131cea685ce |
| ViT-22B | 88 | Not recorded | 68fa74871e209131cea623f3 |
| CoCa-g | 88 | Not recorded | 68fa74871e209131cea623f3 |
| InternVideo2s2-1B | 87.9 | Not recorded | 68fa74871e209131cea623f3 |
| UniFormerV2-L | 87.7 | Not recorded | 68fa75861e209131cea685ce |
| DUALPATH-L | 87.7 | Not recorded | 68fa75861e209131cea685ce |
| VideoPrism-g | 87.2 | Not recorded | 68fa74871e209131cea623f3 |
| ViT-e | 86.5 | Not recorded | 68fa74871e209131cea623f3 |
| MViT-L | 86.1 | Not recorded | 68fa75861e209131cea685ce |
| VideoMAE-L | 86.1 | Not recorded | 68fa75861e209131cea685ce |
| UniFormerV2-B | 85.6 | Not recorded | 68fa75861e209131cea685ce |
| PoseConv3D | 85.5 | Not recorded | 68fa6c321e209131cea32c30 |
| DUALPATH-B | 85.4 | Not recorded | 68fa75861e209131cea685ce |
| VideoMAE1600e | 85.2 | Not recorded | 68fa72b81e209131cea58242 |
| VideoMAE-L | 85.2 | Not recorded | 68fa75861e209131cea685ce |
| VideoSwin | 84.9 | Not recorded | 68fa6c321e209131cea32c30 |
| Swin-L, 384 | 84.9 | Not recorded | 68fa6aa91e209131cea29915 |
| ViViT-H/16×2 | 84.9 | Not recorded | 68fa6aa91e209131cea29915 |
| X-CLIP | 84.7 | Not recorded | 692d2108d6b0dd334293161e |
| UniFormerV2-B | 84.4 | Not recorded | 68fa75861e209131cea685ce |
| ActionCLIP | 83.8 | Not recorded | 692d2108d6b0dd334293161e |
| ActionCLIP | 83.8 | Not recorded | 68fa6aca1e209131cea2a40f |
| R3D-RS-200 (48↑) | 83.5 | Not recorded | 68fa6aa91e209131cea29915 |
| DINOv2-g | 83.4 | Not recorded | 68fa74871e209131cea623f3 |
| VideoSwin-L | 83.1 | Not recorded | 68fa75861e209131cea685ce |
| SMILE | 83.1 | Not recorded | 68fa77b21e209131cea74d3d |
| Swin-L | 83.1 | Not recorded | 692d2108d6b0dd334293161e |
| Swin-L | 83.1 | Not recorded | 68fa6aa91e209131cea29915 |
| Uniformer-B | 83 | Not recorded | 692d2108d6b0dd334293161e |
| UMT-L | 82.8 | Not recorded | 68fa74871e209131cea623f3 |
| ViViT-L/16x2 | 82.8 | Not recorded | 68fa6aca1e209131cea2a40f |
| ST-Adapter | 82.7 | Not recorded | 68fa75861e209131cea685ce |
| ActionCLIP | 82.6 | Not recorded | 68fa6aca1e209131cea2a40f |
| AMD800e | 82.2 | Not recorded | 68fa72b81e209131cea58242 |
| VideoMAEv2-g | 82.1 | Not recorded | 68fa74871e209131cea623f3 |
| V-JEPA-H | 81.9 | Not recorded | 68fa74871e209131cea623f3 |
| ViViT FE | 81.7 | Not recorded | 68fa72b81e209131cea58242 |
| ViViT-L/16×2 FE | 81.7 | Not recorded | 68fa6aa91e209131cea29915 |
| VideoMAE1600e | 81.5 | Not recorded | 68fa72b81e209131cea58242 |
| VideoMAE-B | 81.5 | Not recorded | 68fa75861e209131cea685ce |
| SIGMA | 81.5 | Not recorded | 68fa77b21e209131cea74d3d |
| MME | 81.5 | Not recorded | 68fa77b21e209131cea74d3d |
| MAE-ST | 81.3 | Not recorded | 68fa72b81e209131cea58242 |
| MViTv1-B | 81.2 | Not recorded | 68fa75861e209131cea685ce |
| MGMAE | 81.2 | Not recorded | 68fa77b21e209131cea74d3d |
| MVIT-B | 81.2 | Not recorded | 68fa6aca1e209131cea2a40f |
| DMAE300e | 80.8 | Not recorded | 68fa72b81e209131cea58242 |
| OmniMAE | 80.8 | Not recorded | 68fa77b21e209131cea74d3d |
| MGM | 80.8 | Not recorded | 68fa77b21e209131cea74d3d |
| TimeSformer-L | 80.7 | Not recorded | 68fa6a201e209131cea2621b |
| TimeSformer | 80.7 | Not recorded | 68fa72b81e209131cea58242 |
| TimeSformer-L | 80.7 | Not recorded | 68fa75861e209131cea685ce |
| TimeSformer-L | 80.7 | Not recorded | 692d2108d6b0dd334293161e |
| TimeSformer-L | 80.7 | Not recorded | 68fa6aa91e209131cea29915 |
| TimeSformer-L | 80.7 | Not recorded | 68fa6aca1e209131cea2a40f |
| ViViT-L/16x2 | 80.6 | Not recorded | 68fa6a201e209131cea2621b |
| BEVT | 80.6 | Not recorded | 68fa72b81e209131cea58242 |
| Video Swin | 80.6 | Not recorded | 68fa72b81e209131cea58242 |
| BEVT | 80.6 | Not recorded | 68fa77b21e209131cea74d3d |
| ViViT-L/16x2 | 80.6 | Not recorded | 68fa6aca1e209131cea2a40f |
| STAM | 80.5 | Not recorded | 68fa6aca1e209131cea2a40f |
| X3D-XXL | 80.4 | Not recorded | 68fa6a201e209131cea2621b |
| X3D-XXL | 80.4 | Not recorded | 68fa6aca1e209131cea2a40f |
| X-ViT | 80.2 | Not recorded | 68fa6a201e209131cea2621b |
| Motionformer | 80.2 | Not recorded | 68fa72b81e209131cea58242 |
| VINDLU | 80.1 | Not recorded | 68fa6f081e209131cea436e1 |
| AMD800e | 80.1 | Not recorded | 68fa72b81e209131cea58242 |
| VideoMAE800e | 80 | Not recorded | 68fa72b81e209131cea58242 |
| VideoMAE | 80 | Not recorded | 68fa77b21e209131cea74d3d |
| SlowFast 16x8 R101+NL | 79.8 | Not recorded | 68fa6a201e209131cea2621b |
| SlowFast | 79.8 | Not recorded | 68fa6d001e209131cea37375 |
| SlowFast+NL | 79.8 | Not recorded | 68fa6d2e1e209131cea3821e |
| TVTS (Ours) | 79.8 | Not recorded | 68fa6ec81e209131cea4209d |
| SlowFast | 79.8 | Not recorded | 68fa72b81e209131cea58242 |

### UniFormerV2-L

Top-1 Accuracy: 89.7. Result ID: 6993b9fa3a6c31484918182d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75861e209131cea685ce.md>)

Recorded evaluation configuration: {"Backbone":"ViT-L","Pretraining":"CLIP-400M+K710-0.66M","Views":"32 x 3 x 4","Total Parameters (M)":"354","Trainable Parameters (M)":"354","FLOPs (T)":"16.0"}. This record does not assert independent reproduction.

### AM/12, Transformer

Top-1 Accuracy: 89.6. Result ID: 6993b9fa3a6c314849181833.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75861e209131cea685ce.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75861e209131cea685ce.md>)

Recorded evaluation configuration: {"Backbone":"ViT-B(Dinov2)","Pretraining":"IN-21K","Views":"32 x 3 x 1","Total Parameters (M)":"86+32+45+103","Trainable Parameters (M)":"32+45+103","FLOPs (T)":"13.8"}. This record does not assert independent reproduction.

### MTV-H

Top-1 Accuracy: 89.1. Result ID: 6993b9fa3a6c314849181829.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75861e209131cea685ce.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ca11e209131cea35110.md>)

[Implementation](<https://github.com/google-research/scenic>)

Recorded evaluation configuration: {"Backbone":"ViT-H+B+S+T","Pretraining":"IN-21K+WTS-600M","Views":"32 x 3 x 4","Total Parameters (M)":"1000+","Trainable Parameters (M)":"1000+","FLOPs (T)":"44.5"}. This record does not assert independent reproduction.

### AM/12, Transformer

Top-1 Accuracy: 89.1. Result ID: 6993b9fa3a6c314849181832.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75861e209131cea685ce.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75861e209131cea685ce.md>)

Recorded evaluation configuration: {"Backbone":"ViT-B(Dinov2)","Pretraining":"IN-21K","Views":"8 x 3 x 1","Total Parameters (M)":"86+32+45+103","Trainable Parameters (M)":"32+45+103","FLOPs (T)":"6.9"}. This record does not assert independent reproduction.

### CoCa

Top-1 Accuracy: 88.9. Result ID: 6993b9fa3a6c314849181828.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75861e209131cea685ce.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6cc71e209131cea35ecd.md>)

[Implementation](<https://socraticmodels.github.io/>)

Recorded evaluation configuration: {"Backbone":"ViT-g","Pretraining":"JFT-3B+ALIGN-1.8B","Views":"N/A","Total Parameters (M)":"1000+","Trainable Parameters (M)":"1000+","FLOPs (T)":"N/A"}. This record does not assert independent reproduction.

### InternVideo2s2-6B

Top-1 Accuracy: 88.8. Result ID: 699361a73a6c3148491584a3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74871e209131cea623f3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74871e209131cea623f3.md>)

[Implementation](<https://github.com/OpenGVLab/InternVideo2>)

Recorded evaluation configuration: {"Training Data":"IV-400M","Frames x Resolution":"16 x 224","Probing Type":"Attentive Probing"}. This record does not assert independent reproduction.

### AM/12, LSTM

Top-1 Accuracy: 88.8. Result ID: 6993b9fa3a6c314849181831.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75861e209131cea685ce.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75861e209131cea685ce.md>)

Recorded evaluation configuration: {"Backbone":"ViT-B(Dinov2)","Pretraining":"IN-21K","Views":"8 x 3 x 1","Total Parameters (M)":"86+32+45+360","Trainable Parameters (M)":"32+45+360","FLOPs (T)":"5.3"}. This record does not assert independent reproduction.

### ViT-22B

Top-1 Accuracy: 88. Result ID: 699361a73a6c3148491584a0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74871e209131cea623f3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e901e209131cea40a8d.md>)

Recorded evaluation configuration: {"Training Data":"I-4B","Frames x Resolution":"128 x 224","Probing Type":"Attentive Probing"}. This record does not assert independent reproduction.

### CoCa-g

Top-1 Accuracy: 88. Result ID: 699361a73a6c3148491584a1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74871e209131cea623f3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6cc71e209131cea35ecd.md>)

[Implementation](<https://socraticmodels.github.io/>)

Recorded evaluation configuration: {"Training Data":"I-3B","Frames x Resolution":"16 x 576","Probing Type":"Attentive Probing"}. This record does not assert independent reproduction.

### InternVideo2s2-1B

Top-1 Accuracy: 87.9. Result ID: 699361a73a6c3148491584a2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74871e209131cea623f3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74871e209131cea623f3.md>)

[Implementation](<https://github.com/OpenGVLab/InternVideo2>)

Recorded evaluation configuration: {"Training Data":"IV-25.5M","Frames x Resolution":"16 x 224","Probing Type":"Attentive Probing"}. This record does not assert independent reproduction.

### UniFormerV2-L

Top-1 Accuracy: 87.7. Result ID: 6993b9fa3a6c31484918182b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75861e209131cea685ce.md>)

Recorded evaluation configuration: {"Backbone":"ViT-L","Pretraining":"CLIP-400M","Views":"8 x 3 x 4","Total Parameters (M)":"354","Trainable Parameters (M)":"354","FLOPs (T)":"8.0"}. This record does not assert independent reproduction.

### DUALPATH-L

Top-1 Accuracy: 87.7. Result ID: 6993b9fa3a6c314849181830.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75861e209131cea685ce.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ec71e209131cea4203b.md>)

[Implementation](<https://github.com/park-jungin/DualPath>)

Recorded evaluation configuration: {"Backbone":"ViT-L","Pretraining":"CLIP-400M","Views":"32 x 3 x 1","Total Parameters (M)":"330","Trainable Parameters (M)":"27","FLOPs (T)":"1.9"}. This record does not assert independent reproduction.

### VideoPrism-g

Top-1 Accuracy: 87.2. Result ID: 699361a73a6c31484915849e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74871e209131cea623f3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa792f1e209131cea7c78d.md>)

Recorded evaluation configuration: {"Training Data":"V-619M","Frames x Resolution":"16 x 288","Probing Type":"Attentive Probing"}. This record does not assert independent reproduction.

### ViT-e

Top-1 Accuracy: 86.5. Result ID: 699361a73a6c31484915849f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74871e209131cea623f3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e901e209131cea40a8d.md>)

Recorded evaluation configuration: {"Training Data":"I-4B","Frames x Resolution":"128 x 224","Probing Type":"Attentive Probing"}. This record does not assert independent reproduction.

### MViT-L

Top-1 Accuracy: 86.1. Result ID: 6993b9fa3a6c314849181823.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75861e209131cea685ce.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c2f1e209131cea32a41.md>)

[Implementation](<https://github.com/facebookresearch/mvit>)

Recorded evaluation configuration: {"Backbone":"MViTv2-L","Pretraining":"IN-21K","Views":"40 x 3 x 5","Total Parameters (M)":"218","Trainable Parameters (M)":"218","FLOPs (T)":"42.4"}. This record does not assert independent reproduction.

### VideoMAE-L

Top-1 Accuracy: 86.1. Result ID: 6993b9fa3a6c314849181826.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75861e209131cea685ce.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6da81e209131cea3b2fd.md>)

[Implementation](<https://github.com/MCG-NJU/VideoMAE>)

Recorded evaluation configuration: {"Backbone":"ViT-L","Pretraining":"N/A","Views":"40 x 3 x 4","Total Parameters (M)":"305","Trainable Parameters (M)":"305","FLOPs (T)":"47.5"}. This record does not assert independent reproduction.

### UniFormerV2-B

Top-1 Accuracy: 85.6. Result ID: 6993b9fa3a6c31484918182c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75861e209131cea685ce.md>)

Recorded evaluation configuration: {"Backbone":"ViT-B","Pretraining":"CLIP-400M+K710-0.66M","Views":"8 x 3 x 4","Total Parameters (M)":"115","Trainable Parameters (M)":"115","FLOPs (T)":"1.6"}. This record does not assert independent reproduction.

### PoseConv3D

Top-1 Accuracy: 85.5. Result ID: 698ef1fe3a6c314849f3bf34.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c321e209131cea32c30.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c321e209131cea32c30.md>)

[Implementation](<https://github.com/kennymckormick/pyskl>)

Recorded evaluation configuration: {"Modality":"RGB + Pose","Fusion":"LateFusion"}. This record does not assert independent reproduction.

### DUALPATH-B

Top-1 Accuracy: 85.4. Result ID: 6993b9fa3a6c31484918182f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75861e209131cea685ce.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ec71e209131cea4203b.md>)

[Implementation](<https://github.com/park-jungin/DualPath>)

Recorded evaluation configuration: {"Backbone":"ViT-B","Pretraining":"CLIP-400M","Views":"32 x 3 x 1","Total Parameters (M)":"96","Trainable Parameters (M)":"10","FLOPs (T)":"0.7"}. This record does not assert independent reproduction.

### VideoMAE1600e

Top-1 Accuracy: 85.2. Result ID: 6992552b3a6c3148490e3ea5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72b81e209131cea58242.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6da81e209131cea3b2fd.md>)

[Implementation](<https://github.com/MCG-NJU/VideoMAE>)

Recorded evaluation configuration: {"Backbone":"ViT-L","Extra data":null,"Extra labels":false,"Frames":16,"GFLOPS":"597×5×3","Param":"305M"}. This record does not assert independent reproduction.

### VideoMAE-L

Top-1 Accuracy: 85.2. Result ID: 6993b9fa3a6c314849181825.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75861e209131cea685ce.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6da81e209131cea3b2fd.md>)

[Implementation](<https://github.com/MCG-NJU/VideoMAE>)

Recorded evaluation configuration: {"Backbone":"ViT-L","Pretraining":"N/A","Views":"16 x 3 x 5","Total Parameters (M)":"305","Trainable Parameters (M)":"305","FLOPs (T)":"9.0"}. This record does not assert independent reproduction.

### VideoSwin

Top-1 Accuracy: 84.9. Result ID: 698ef1fe3a6c314849f3bf32.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6c321e209131cea32c30.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a391e209131cea26be4.md>)

[Implementation](<https://github.com/SwinTransformer/Video-Swin-Transformer>)

Recorded evaluation configuration: {"Modality":"RGB"}. This record does not assert independent reproduction.

### Swin-L, 384

Top-1 Accuracy: 84.9. Result ID: 699ac9df5698665ec4585fd5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6aa91e209131cea29915.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a391e209131cea26be4.md>)

[Implementation](<https://github.com/SwinTransformer/Video-Swin-Transformer>)

Recorded evaluation configuration: {"Pretrain":"ImageNet-21K","FLOPs (B) × Views":"2107 × 10 × 5"}. This record does not assert independent reproduction.

### ViViT-H/16×2

Top-1 Accuracy: 84.9. Result ID: 699ac9df5698665ec4585fd6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6aa91e209131cea29915.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b2d1e209131cea2c9cc.md>)

[Implementation](<https://github.com/google-research/scenic>)

Recorded evaluation configuration: {"Pretrain":"JFT","FLOPs (B) × Views":"3981 × 4 × 3"}. This record does not assert independent reproduction.

### X-CLIP

Top-1 Accuracy: 84.7. Result ID: 6995e71d3badfaffe6ef52f6.

[Reporting paper](<https://www.sota2.com/research/paper/692d2108d6b0dd334293161e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d331e209131cea385df.md>)

[Implementation](<https://www.sota2.com/aka.ms/X-CLIP>)

Recorded evaluation configuration: {"Frames":16,"Views":"4 x 3","Protocol":"Fully-supervised"}. This record does not assert independent reproduction.

### UniFormerV2-B

Top-1 Accuracy: 84.4. Result ID: 6993b9fa3a6c31484918182a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75861e209131cea685ce.md>)

Recorded evaluation configuration: {"Backbone":"ViT-B","Pretraining":"CLIP-400M","Views":"8 x 3 x 4","Total Parameters (M)":"115","Trainable Parameters (M)":"115","FLOPs (T)":"1.6"}. This record does not assert independent reproduction.

### ActionCLIP

Top-1 Accuracy: 83.8. Result ID: 6995e71d3badfaffe6ef52f5.

[Reporting paper](<https://www.sota2.com/research/paper/692d2108d6b0dd334293161e.md>)

Recorded evaluation configuration: {"Frames":32,"Views":"10 x 3","Protocol":"Fully-supervised"}. This record does not assert independent reproduction.

### ActionCLIP

Top-1 Accuracy: 83.8. Result ID: 699ad1405698665ec4603c71.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6aca1e209131cea2a40f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6aca1e209131cea2a40f.md>)

[Implementation](<https://github.com/sallymmx/ActionCLIP.git>)

Recorded evaluation configuration: {"Backbone":"ViT-B/16","Frames":"32"}. This record does not assert independent reproduction.

### R3D-RS-200 (48↑)

Top-1 Accuracy: 83.5. Result ID: 699ac9df5698665ec4585fd7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6aa91e209131cea29915.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6aa91e209131cea29915.md>)

[Implementation](<https://github.com/tensorflow/models/tree/master/official>)

Recorded evaluation configuration: {"Pretrain":"WVT","FLOPs (B) × Views":"307 × 10 × 3"}. This record does not assert independent reproduction.

### DINOv2-g

Top-1 Accuracy: 83.4. Result ID: 699361a73a6c31484915849d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74871e209131cea623f3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72181e209131cea546c3.md>)

[Implementation](<https://github.com/facebookresearch/dinov2>)

Recorded evaluation configuration: {"Training Data":"I-142M","Frames x Resolution":"16 x 224","Probing Type":"Attentive Probing"}. This record does not assert independent reproduction.

### VideoSwin-L

Top-1 Accuracy: 83.1. Result ID: 6993b9fa3a6c314849181822.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75861e209131cea685ce.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a391e209131cea26be4.md>)

[Implementation](<https://github.com/SwinTransformer/Video-Swin-Transformer>)

Recorded evaluation configuration: {"Backbone":"Swin-L","Pretraining":"IN-21K","Views":"32 x 3 x 4","Total Parameters (M)":"197","Trainable Parameters (M)":"197","FLOPs (T)":"7.2"}. This record does not assert independent reproduction.

### SMILE

Top-1 Accuracy: 83.1. Result ID: 69955f273badfaffe6513e20.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77b21e209131cea74d3d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77b21e209131cea74d3d.md>)

[Implementation](<https://github.com/fmthoker/SMILE>)

Recorded evaluation configuration: {"Backbone":"ViT-B","Epochs":600,"Pretrain":"K400","Evaluation Protocol":"Full finetuning"}. This record does not assert independent reproduction.

### Swin-L

Top-1 Accuracy: 83.1. Result ID: 6995e71d3badfaffe6ef52f4.

[Reporting paper](<https://www.sota2.com/research/paper/692d2108d6b0dd334293161e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a391e209131cea26be4.md>)

[Implementation](<https://github.com/SwinTransformer/Video-Swin-Transformer>)

Recorded evaluation configuration: {"Frames":32,"Views":"4 x 3","Protocol":"Fully-supervised"}. This record does not assert independent reproduction.

### Swin-L

Top-1 Accuracy: 83.1. Result ID: 699ac9df5698665ec4585fd4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6aa91e209131cea29915.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a391e209131cea26be4.md>)

[Implementation](<https://github.com/SwinTransformer/Video-Swin-Transformer>)

Recorded evaluation configuration: {"Pretrain":"ImageNet-21K","FLOPs (B) × Views":"604 × 4 × 3"}. This record does not assert independent reproduction.

### Uniformer-B

Top-1 Accuracy: 83. Result ID: 6995e71d3badfaffe6ef52f2.

[Reporting paper](<https://www.sota2.com/research/paper/692d2108d6b0dd334293161e.md>)

Recorded evaluation configuration: {"Frames":32,"Views":"4 x 3","Protocol":"Fully-supervised"}. This record does not assert independent reproduction.

### UMT-L

Top-1 Accuracy: 82.8. Result ID: 699361a73a6c31484915849a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74871e209131cea623f3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72711e209131cea5652b.md>)

[Implementation](<https://github.com/OpenGVLab/unmasked_teacher>)

Recorded evaluation configuration: {"Training Data":"IV-25M","Probing Type":"Attentive Probing"}. This record does not assert independent reproduction.

### ViViT-L/16x2

Top-1 Accuracy: 82.8. Result ID: 699ad1405698665ec4603c6f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6aca1e209131cea2a40f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b2d1e209131cea2c9cc.md>)

[Implementation](<https://github.com/google-research/scenic>)

Recorded evaluation configuration: {"Frames":"32","Pre-training":"JFT"}. This record does not assert independent reproduction.

### ST-Adapter

Top-1 Accuracy: 82.7. Result ID: 6993b9fa3a6c31484918182e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75861e209131cea685ce.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d9f1e209131cea3aeec.md>)

[Implementation](<https://github.com/linziyi96/st-adapter>)

Recorded evaluation configuration: {"Backbone":"ViT-B","Pretraining":"CLIP-400M","Views":"32 x 3 x 1","Total Parameters (M)":"93","Trainable Parameters (M)":"7","FLOPs (T)":"1.8"}. This record does not assert independent reproduction.

### ActionCLIP

Top-1 Accuracy: 82.6. Result ID: 699ad1405698665ec4603c70.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6aca1e209131cea2a40f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6aca1e209131cea2a40f.md>)

[Implementation](<https://github.com/sallymmx/ActionCLIP.git>)

Recorded evaluation configuration: {"Backbone":"ViT-B/16","Frames":"16"}. This record does not assert independent reproduction.

### AMD800e

Top-1 Accuracy: 82.2. Result ID: 6992552b3a6c3148490e3e9e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72b81e209131cea58242.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72b81e209131cea58242.md>)

[Implementation](<https://github.com/MCG-NJU/VideoMAE>)

Recorded evaluation configuration: {"Backbone":"ViT-B","Extra data":null,"Extra labels":false,"Frames":16,"GFLOPS":"180×5×3","Param":"87M"}. This record does not assert independent reproduction.

### VideoMAEv2-g

Top-1 Accuracy: 82.1. Result ID: 699361a73a6c31484915849b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74871e209131cea623f3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f171e209131cea43db1.md>)

Recorded evaluation configuration: {"Training Data":"V-1.35M","Probing Type":"Attentive Probing"}. This record does not assert independent reproduction.

### V-JEPA-H

Top-1 Accuracy: 81.9. Result ID: 699361a73a6c31484915849c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74871e209131cea623f3.md>)

Recorded evaluation configuration: {"Training Data":"V-2M","Frames x Resolution":"16 x 384","Probing Type":"Attentive Probing"}. This record does not assert independent reproduction.

### ViViT FE

Top-1 Accuracy: 81.7. Result ID: 6992552b3a6c3148490e3ea1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72b81e209131cea58242.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b2d1e209131cea2c9cc.md>)

[Implementation](<https://github.com/google-research/scenic>)

Recorded evaluation configuration: {"Backbone":"ViT-L","Extra data":"ImageNet-21K","Extra labels":true,"Frames":128,"GFLOPS":"3980×1×3","Param":null}. This record does not assert independent reproduction.

### ViViT-L/16×2 FE

Top-1 Accuracy: 81.7. Result ID: 699ac9df5698665ec4585fd3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6aa91e209131cea29915.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b2d1e209131cea2c9cc.md>)

[Implementation](<https://github.com/google-research/scenic>)

Recorded evaluation configuration: {"Pretrain":"ImageNet-21K","FLOPs (B) × Views":"3980 × 1 × 3"}. This record does not assert independent reproduction.

### VideoMAE1600e

Top-1 Accuracy: 81.5. Result ID: 6992552b3a6c3148490e3e99.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72b81e209131cea58242.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6da81e209131cea3b2fd.md>)

[Implementation](<https://github.com/MCG-NJU/VideoMAE>)

Recorded evaluation configuration: {"Backbone":"ViT-B","Extra data":null,"Extra labels":false,"Frames":16,"GFLOPS":"180×5×3","Param":"87M"}. This record does not assert independent reproduction.

### VideoMAE-B

Top-1 Accuracy: 81.5. Result ID: 6993b9fa3a6c314849181824.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75861e209131cea685ce.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6da81e209131cea3b2fd.md>)

[Implementation](<https://github.com/MCG-NJU/VideoMAE>)

Recorded evaluation configuration: {"Backbone":"ViT-B","Pretraining":"N/A","Views":"16 x 3 x 5","Total Parameters (M)":"87","Trainable Parameters (M)":"87","FLOPs (T)":"2.7"}. This record does not assert independent reproduction.

### SIGMA

Top-1 Accuracy: 81.5. Result ID: 69955f273badfaffe6513e1e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77b21e209131cea74d3d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa742d1e209131cea60408.md>)

Recorded evaluation configuration: {"Backbone":"ViT-B","Epochs":800,"Pretrain":"K400","Evaluation Protocol":"Full finetuning"}. This record does not assert independent reproduction.

### MME

Top-1 Accuracy: 81.5. Result ID: 69955f273badfaffe6513e1f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77b21e209131cea74d3d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ece1e209131cea42450.md>)

[Implementation](<https://github.com/XinyuSun/MME>)

Recorded evaluation configuration: {"Backbone":"ViT-B","Epochs":800,"Pretrain":"K400","Evaluation Protocol":"Full finetuning","evaluation\_note":"results obtained by authors' evaluation"}. This record does not assert independent reproduction.

### MAE-ST

Top-1 Accuracy: 81.3. Result ID: 6992552b3a6c3148490e3e97.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72b81e209131cea58242.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6dbe1e209131cea3bcb2.md>)

[Implementation](<https://github.com/facebookresearch/mae_st>)

Recorded evaluation configuration: {"Backbone":"ViT-B","Extra data":null,"Extra labels":false,"Frames":16,"GFLOPS":"180×7×3","Param":"87M"}. This record does not assert independent reproduction.

### MViTv1-B

Top-1 Accuracy: 81.2. Result ID: 6993b9fa3a6c314849181821.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75861e209131cea685ce.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69cb1e209131cea2412c.md>)

[Implementation](<https://github.com/facebookresearch/SlowFast>)

Recorded evaluation configuration: {"Backbone":"MViTv1-B","Pretraining":"N/A","Views":"64 x 3 x 3","Total Parameters (M)":"37","Trainable Parameters (M)":"37","FLOPs (T)":"4.1"}. This record does not assert independent reproduction.

### MGMAE

Top-1 Accuracy: 81.2. Result ID: 69955f273badfaffe6513e1d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77b21e209131cea74d3d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70621e209131cea4ac7a.md>)

Recorded evaluation configuration: {"Backbone":"ViT-B","Epochs":800,"Pretrain":"K400","Evaluation Protocol":"Full finetuning"}. This record does not assert independent reproduction.

### MVIT-B

Top-1 Accuracy: 81.2. Result ID: 699ad1405698665ec4603c6b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6aca1e209131cea2a40f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69cb1e209131cea2412c.md>)

[Implementation](<https://github.com/facebookresearch/SlowFast>)

Recorded evaluation configuration: {"Frames":"64"}. This record does not assert independent reproduction.

### DMAE300e

Top-1 Accuracy: 80.8. Result ID: 6992552b3a6c3148490e3e9d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72b81e209131cea58242.md>)

Recorded evaluation configuration: {"Backbone":"ViT-B","Extra data":null,"Extra labels":false,"Frames":16,"GFLOPS":"180×5×3","Param":"87M","is\_our\_implementation":true}. This record does not assert independent reproduction.

### OmniMAE

Top-1 Accuracy: 80.8. Result ID: 69955f273badfaffe6513e1b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77b21e209131cea74d3d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f961e209131cea4674e.md>)

[Implementation](<https://github.com/facebookresearch/omnivore>)

Recorded evaluation configuration: {"Backbone":"ViT-B","Epochs":800,"Pretrain":"K400","Evaluation Protocol":"Full finetuning"}. This record does not assert independent reproduction.

### MGM

Top-1 Accuracy: 80.8. Result ID: 69955f273badfaffe6513e1c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77b21e209131cea74d3d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70601e209131cea4ab1d.md>)

Recorded evaluation configuration: {"Backbone":"ViT-B","Epochs":800,"Pretrain":"K400","Evaluation Protocol":"Full finetuning"}. This record does not assert independent reproduction.

### TimeSformer-L

Top-1 Accuracy: 80.7. Result ID: 698e11dbffc3901de09a902e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6a201e209131cea2621b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a1d1e209131cea25fbb.md>)

[Implementation](<https://github.com/facebookresearch/TimeSformer>)

Recorded evaluation configuration: {"Views":"1 x 3","FLOPs (x10^9)":7140}. This record does not assert independent reproduction.

### TimeSformer

Top-1 Accuracy: 80.7. Result ID: 6992552b3a6c3148490e3ea4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72b81e209131cea58242.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a1d1e209131cea25fbb.md>)

[Implementation](<https://github.com/facebookresearch/TimeSformer>)

Recorded evaluation configuration: {"Backbone":"ViT-L","Extra data":"ImageNet-21K","Extra labels":true,"Frames":96,"GFLOPS":"8353×1×3","Param":"430M"}. This record does not assert independent reproduction.

### TimeSformer-L

Top-1 Accuracy: 80.7. Result ID: 6993b9fa3a6c314849181827.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75861e209131cea685ce.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a1d1e209131cea25fbb.md>)

[Implementation](<https://github.com/facebookresearch/TimeSformer>)

Recorded evaluation configuration: {"Backbone":"ViT-B","Pretraining":"IN-21K","Views":"96 x 3 x 1","Total Parameters (M)":"121","Trainable Parameters (M)":"121","FLOPs (T)":"7.1"}. This record does not assert independent reproduction.

### TimeSformer-L

Top-1 Accuracy: 80.7. Result ID: 6995e71d3badfaffe6ef52f3.

[Reporting paper](<https://www.sota2.com/research/paper/692d2108d6b0dd334293161e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a1d1e209131cea25fbb.md>)

[Implementation](<https://github.com/facebookresearch/TimeSformer>)

Recorded evaluation configuration: {"Frames":96,"Views":"1 x 3","Protocol":"Fully-supervised"}. This record does not assert independent reproduction.

### TimeSformer-L

Top-1 Accuracy: 80.7. Result ID: 699ac9df5698665ec4585fd2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6aa91e209131cea29915.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a1d1e209131cea25fbb.md>)

[Implementation](<https://github.com/facebookresearch/TimeSformer>)

Recorded evaluation configuration: {"Pretrain":"ImageNet-21K","FLOPs (B) × Views":"2380 × 1 × 3"}. This record does not assert independent reproduction.

### TimeSformer-L

Top-1 Accuracy: 80.7. Result ID: 699ad1405698665ec4603c6d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6aca1e209131cea2a40f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a1d1e209131cea25fbb.md>)

[Implementation](<https://github.com/facebookresearch/TimeSformer>)

Recorded evaluation configuration: {"Frames":"96"}. This record does not assert independent reproduction.

### ViViT-L/16x2

Top-1 Accuracy: 80.6. Result ID: 698e11dbffc3901de09a902f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6a201e209131cea2621b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b2d1e209131cea2c9cc.md>)

[Implementation](<https://github.com/google-research/scenic>)

Recorded evaluation configuration: {"Views":"4 x 3","FLOPs (x10^9)":17352}. This record does not assert independent reproduction.

### BEVT

Top-1 Accuracy: 80.6. Result ID: 6992552b3a6c3148490e3e9b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72b81e209131cea58242.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6bf51e209131cea3138e.md>)

[Implementation](<https://github.com/xyzforever/BEVT>)

Recorded evaluation configuration: {"Backbone":"Swin-B","Extra data":"IN-1K+DALLE","Extra labels":false,"Frames":32,"GFLOPS":"282×4×3","Param":"88M"}. This record does not assert independent reproduction.

### Video Swin

Top-1 Accuracy: 80.6. Result ID: 6992552b3a6c3148490e3e9c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72b81e209131cea58242.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a391e209131cea26be4.md>)

[Implementation](<https://github.com/SwinTransformer/Video-Swin-Transformer>)

Recorded evaluation configuration: {"Backbone":"Swin-B","Extra data":"ImageNet-21K","Extra labels":true,"Frames":32,"GFLOPS":"282×4×3","Param":"88M"}. This record does not assert independent reproduction.

### BEVT

Top-1 Accuracy: 80.6. Result ID: 69955f273badfaffe6513e1a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77b21e209131cea74d3d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6bf51e209131cea3138e.md>)

[Implementation](<https://github.com/xyzforever/BEVT>)

Recorded evaluation configuration: {"Backbone":"ViT-B","Epochs":null,"Pretrain":"K400","Evaluation Protocol":"Full finetuning"}. This record does not assert independent reproduction.

### ViViT-L/16x2

Top-1 Accuracy: 80.6. Result ID: 699ad1405698665ec4603c6e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6aca1e209131cea2a40f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b2d1e209131cea2c9cc.md>)

[Implementation](<https://github.com/google-research/scenic>)

Recorded evaluation configuration: {"Frames":"32"}. This record does not assert independent reproduction.

### STAM

Top-1 Accuracy: 80.5. Result ID: 699ad1405698665ec4603c6c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6aca1e209131cea2a40f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a021e209131cea255f2.md>)

[Implementation](<https://github.com/Alibaba-MIIL/STAM>)

Recorded evaluation configuration: {"Frames":"64"}. This record does not assert independent reproduction.

### X3D-XXL

Top-1 Accuracy: 80.4. Result ID: 698e11dbffc3901de09a902d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6a201e209131cea2621b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67531e209131cea15826.md>)

Recorded evaluation configuration: {"Views":"10 x 3","FLOPs (x10^9)":5823}. This record does not assert independent reproduction.

### X3D-XXL

Top-1 Accuracy: 80.4. Result ID: 699ad1405698665ec4603c5e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6aca1e209131cea2a40f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67531e209131cea15826.md>)

Recorded evaluation configuration: {"Frames":"16"}. This record does not assert independent reproduction.

### X-ViT

Top-1 Accuracy: 80.2. Result ID: 698e11dbffc3901de09a9031.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6a201e209131cea2621b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a201e209131cea2621b.md>)

Recorded evaluation configuration: {"Frames":"16x","Views":"1 x 3","FLOPs (x10^9)":850}. This record does not assert independent reproduction.

### Motionformer

Top-1 Accuracy: 80.2. Result ID: 6992552b3a6c3148490e3ea3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72b81e209131cea58242.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b191e209131cea2c2a8.md>)

[Implementation](<https://github.com/facebookresearch/Motionformer>)

Recorded evaluation configuration: {"Backbone":"ViT-L","Extra data":"ImageNet-21K","Extra labels":true,"Frames":32,"GFLOPS":"1185×10×3","Param":"382M"}. This record does not assert independent reproduction.

### VINDLU

Top-1 Accuracy: 80.1. Result ID: 69907d843a6c31484900ee35.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f081e209131cea436e1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f081e209131cea436e1.md>)

[Implementation](<https://github.com/klauscc/VindLU>)

Recorded evaluation configuration: {"Backbone":"TimeSformer"}. This record does not assert independent reproduction.

### AMD800e

Top-1 Accuracy: 80.1. Result ID: 6992552b3a6c3148490e3e94.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72b81e209131cea58242.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72b81e209131cea58242.md>)

[Implementation](<https://github.com/MCG-NJU/VideoMAE>)

Recorded evaluation configuration: {"Backbone":"ViT-S","Extra data":null,"Extra labels":false,"Frames":16,"GFLOPS":"57×5×3","Param":"22M"}. This record does not assert independent reproduction.

### VideoMAE800e

Top-1 Accuracy: 80. Result ID: 6992552b3a6c3148490e3e98.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72b81e209131cea58242.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6da81e209131cea3b2fd.md>)

[Implementation](<https://github.com/MCG-NJU/VideoMAE>)

Recorded evaluation configuration: {"Backbone":"ViT-B","Extra data":null,"Extra labels":false,"Frames":16,"GFLOPS":"180×5×3","Param":"87M"}. This record does not assert independent reproduction.

### VideoMAE

Top-1 Accuracy: 80. Result ID: 69955f273badfaffe6513e19.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77b21e209131cea74d3d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6da81e209131cea3b2fd.md>)

[Implementation](<https://github.com/MCG-NJU/VideoMAE>)

Recorded evaluation configuration: {"Backbone":"ViT-B","Epochs":800,"Pretrain":"K400","Evaluation Protocol":"Full finetuning"}. This record does not assert independent reproduction.

### SlowFast 16x8 R101+NL

Top-1 Accuracy: 79.8. Result ID: 698e11dbffc3901de09a902c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6a201e209131cea2621b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66911e209131cea10f0f.md>)

[Implementation](<https://github.com/facebookresearch/SlowFast>)

Recorded evaluation configuration: {"Views":"10 x 3","FLOPs (x10^9)":7020}. This record does not assert independent reproduction.

### SlowFast

Top-1 Accuracy: 79.8. Result ID: 698f77423a6c314849f8d4e5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d001e209131cea37375.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66911e209131cea10f0f.md>)

[Implementation](<https://github.com/facebookresearch/SlowFast>)

Recorded evaluation configuration: {"Pretrain":"Scratch","Frame":"16+64","FLOPS x Views":"234G × 30"}. This record does not assert independent reproduction.

### SlowFast+NL

Top-1 Accuracy: 79.8. Result ID: 698f95ae3a6c314849fa4be6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d2e1e209131cea3821e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66911e209131cea10f0f.md>)

[Implementation](<https://github.com/facebookresearch/SlowFast>)

Recorded evaluation configuration: {"Backbone":"3D-ResNet-101","Pre-train":"None","Frames":"(16+64)x30","GFLOPS":"234x30"}. This record does not assert independent reproduction.

### TVTS (Ours)

Top-1 Accuracy: 79.8. Result ID: 699040e23a6c314849ff405d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6ec81e209131cea4209d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ec81e209131cea4209d.md>)

[Implementation](<https://github.com/TencentARC/TVTS>)

Recorded evaluation configuration: {"Backbone":"ViT-B","Pre-train Dataset":"YT-Temporal, CC3M, WebVid2M","Fine-tuning":"true"}. This record does not assert independent reproduction.

### SlowFast

Top-1 Accuracy: 79.8. Result ID: 6992552b3a6c3148490e3e96.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72b81e209131cea58242.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66911e209131cea10f0f.md>)

[Implementation](<https://github.com/facebookresearch/SlowFast>)

Recorded evaluation configuration: {"Backbone":"R101+NL","Extra data":null,"Extra labels":false,"Frames":"16+64","GFLOPS":"234×10×3","Param":"60M"}. This record does not assert independent reproduction.

Page 1 of 4. This is an index page, not the full collection.

- [Next index page](<https://www.sota2.com/research/sota/action-recognition-on-kinetics-400-test.md?metric=Top-1%20Accuracy&page=2>): Follow this link to discover more pages.
