# Image Classification on ImageNet 1K (val) (Speedup) Source: [Image Classification on ImageNet 1K (val) (Speedup)]() Collection index: [llms.txt]() Data updated: 2026-04-20T06:53:51.939Z Evaluation of Image Classification inference speed or efficiency on the validation split of ImageNet 1K, reporting Speedup relative to a baseline. 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 | Image Classification | | Dataset (including any stated split) | ImageNet 1K (val) | | Metric | Speedup | | Unit | SCALAR | | Direction | Higher is better | | Results loaded | 16 of 16 | | 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 - [Speedup](): Higher is better; unit: SCALAR. ## Recorded results | Method | Speedup | Result date | Report paper | | --- | --- | --- | --- | | Threshold-ℓ2 | 2.13 | Not recorded | 69e5b3057d0172140ec7df43 | | DynamicViT | 2.06 | Not recorded | 69e5b3057d0172140ec7df43 | | EViT | 1.94 | Not recorded | 69e5b3057d0172140ec7df43 | | ATS | 1.65 | Not recorded | 69e5b3057d0172140ec7df43 | | Threshold-ℓ2 | 1.37 | Not recorded | 69e5b3057d0172140ec7df43 | | DynamicViT | 1.33 | Not recorded | 69e5b3057d0172140ec7df43 | | EViT | 1.28 | Not recorded | 69e5b3057d0172140ec7df43 | | ATS | 1.14 | Not recorded | 69e5b3057d0172140ec7df43 | | Threshold-ℓ2 | 0.62 | Not recorded | 69e5b3057d0172140ec7df43 | | Threshold-ℓ2 | 0.62 | Not recorded | 69e5b3057d0172140ec7df43 | | DynamicViT | 0.61 | Not recorded | 69e5b3057d0172140ec7df43 | | DynamicViT | 0.61 | Not recorded | 69e5b3057d0172140ec7df43 | | EViT | 0.6 | Not recorded | 69e5b3057d0172140ec7df43 | | EViT | 0.6 | Not recorded | 69e5b3057d0172140ec7df43 | | ATS | 0.57 | Not recorded | 69e5b3057d0172140ec7df43 | | ATS | 0.57 | Not recorded | 69e5b3057d0172140ec7df43 | ### Threshold-ℓ2 Speedup: 2.13. Result ID: 69e5ccb847f34ce924cb68ce. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backend":"Triton (ours)","Token Pruning Ratio":"90%"}. This record does not assert independent reproduction. ### DynamicViT Speedup: 2.06. Result ID: 69e5ccb847f34ce924cb68d2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backend":"Triton (ours)","Token Pruning Ratio":"90%"}. This record does not assert independent reproduction. ### EViT Speedup: 1.94. Result ID: 69e5ccb847f34ce924cb68d6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backend":"Triton (ours)","Token Pruning Ratio":"90%"}. This record does not assert independent reproduction. ### ATS Speedup: 1.65. Result ID: 69e5ccb847f34ce924cb68da. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backend":"Triton (ours)","Token Pruning Ratio":"90%"}. This record does not assert independent reproduction. ### Threshold-ℓ2 Speedup: 1.37. Result ID: 69e5ccb847f34ce924cb68cd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backend":"Triton (ours)","Token Pruning Ratio":"50%"}. This record does not assert independent reproduction. ### DynamicViT Speedup: 1.33. Result ID: 69e5ccb847f34ce924cb68d1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backend":"Triton (ours)","Token Pruning Ratio":"50%"}. This record does not assert independent reproduction. ### EViT Speedup: 1.28. Result ID: 69e5ccb847f34ce924cb68d5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backend":"Triton (ours)","Token Pruning Ratio":"50%"}. This record does not assert independent reproduction. ### ATS Speedup: 1.14. Result ID: 69e5ccb847f34ce924cb68d9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backend":"Triton (ours)","Token Pruning Ratio":"50%"}. This record does not assert independent reproduction. ### Threshold-ℓ2 Speedup: 0.62. Result ID: 69e5ccb847f34ce924cb68cb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backend":"PyTorch (padded)","Token Pruning Ratio":"50%"}. This record does not assert independent reproduction. ### Threshold-ℓ2 Speedup: 0.62. Result ID: 69e5ccb847f34ce924cb68cc. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backend":"PyTorch (padded)","Token Pruning Ratio":"90%"}. This record does not assert independent reproduction. ### DynamicViT Speedup: 0.61. Result ID: 69e5ccb847f34ce924cb68cf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backend":"PyTorch (padded)","Token Pruning Ratio":"50%"}. This record does not assert independent reproduction. ### DynamicViT Speedup: 0.61. Result ID: 69e5ccb847f34ce924cb68d0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backend":"PyTorch (padded)","Token Pruning Ratio":"90%"}. This record does not assert independent reproduction. ### EViT Speedup: 0.6. Result ID: 69e5ccb847f34ce924cb68d3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backend":"PyTorch (padded)","Token Pruning Ratio":"50%"}. This record does not assert independent reproduction. ### EViT Speedup: 0.6. Result ID: 69e5ccb847f34ce924cb68d4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backend":"PyTorch (padded)","Token Pruning Ratio":"90%"}. This record does not assert independent reproduction. ### ATS Speedup: 0.57. Result ID: 69e5ccb847f34ce924cb68d7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backend":"PyTorch (padded)","Token Pruning Ratio":"50%"}. This record does not assert independent reproduction. ### ATS Speedup: 0.57. Result ID: 69e5ccb847f34ce924cb68d8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backend":"PyTorch (padded)","Token Pruning Ratio":"90%"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.