# Image Classification on ImageNet (test) (Step-based Accuracy) Source: [Image Classification on ImageNet (test) (Step-based Accuracy)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:22:34.818Z Evaluation of Image Classification performance on the test split of ImageNet, reporting accuracy metrics measured across various step sizes (e.g., 32, 128, 256 steps). 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 (test) | | Metric | ANN Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 9 of 9 | | 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 - [ANN Accuracy](): Higher is better; unit: PERCENT. - [Accuracy (32 steps)](): Higher is better; unit: PERCENT. - [Accuracy (64 steps)](): Higher is better; unit: PERCENT. - [Accuracy (128 steps)](): Higher is better; unit: PERCENT. - [Accuracy (256 steps)](): Higher is better; unit: PERCENT. - [Accuracy (\>=512 steps)](): Higher is better; unit: PERCENT. ## Recorded results | Method | ANN Accuracy | Result date | Report paper | | --- | --- | --- | --- | | LCP | 75.66 | Not recorded | 68fa77481e209131cea72386 | | ACP | 75.66 | Not recorded | 68fa77481e209131cea72386 | | LCP | 73.4 | Not recorded | 68fa77481e209131cea72386 | | Local threshold balancing (Ours) | 73.31 | Not recorded | 68fa77481e209131cea72386 | | QCFS | 73.3 | Not recorded | 68fa77481e209131cea72386 | | Local threshold balancing (Ours\*) | 73.3 | Not recorded | 68fa77481e209131cea72386 | | Local threshold balancing (Ours) | 71.88 | Not recorded | 68fa77481e209131cea72386 | | RMP | 70.64 | Not recorded | 68fa77481e209131cea72386 | | Local threshold balancing (Ours) | 69.76 | Not recorded | 68fa77481e209131cea72386 | ### LCP ANN Accuracy: 75.66. Result ID: 69947df63a6c3148491d9774. [Reporting paper]() Recorded evaluation configuration: {"Architecture":"ResNet-34 (with 2 extra layers in first conv-block)","Evaluation Protocol":"Post-Training"}. This record does not assert independent reproduction. ### ACP ANN Accuracy: 75.66. Result ID: 69947df63a6c3148491d977a. [Reporting paper]() Recorded evaluation configuration: {"Architecture":"ResNet-34 (with 2 extra layers in first conv-block)","Evaluation Protocol":"Training-involved"}. This record does not assert independent reproduction. ### LCP ANN Accuracy: 73.4. Result ID: 69947df63a6c3148491d9775. [Reporting paper]() Recorded evaluation configuration: {"Architecture":"MobileNet","Evaluation Protocol":"Post-Training"}. This record does not assert independent reproduction. ### Local threshold balancing (Ours) ANN Accuracy: 73.31. Result ID: 69947df63a6c3148491d9778. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Architecture":"ResNet-34","Evaluation Protocol":"Post-Training"}. This record does not assert independent reproduction. ### QCFS ANN Accuracy: 73.3. Result ID: 69947df63a6c3148491d9779. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Architecture":"ResNet-34","Evaluation Protocol":"Training-involved"}. This record does not assert independent reproduction. ### Local threshold balancing (Ours\*) ANN Accuracy: 73.3. Result ID: 69947df63a6c3148491d977b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Architecture":"ResNet-34","Evaluation Protocol":"Training-involved"}. This record does not assert independent reproduction. ### Local threshold balancing (Ours) ANN Accuracy: 71.88. Result ID: 69947df63a6c3148491d9776. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Architecture":"MobileNet V2","Evaluation Protocol":"Post-Training"}. This record does not assert independent reproduction. ### RMP ANN Accuracy: 70.64. Result ID: 69947df63a6c3148491d9773. [Reporting paper]() Recorded evaluation configuration: {"Architecture":"ResNet-34 (without BN)","Evaluation Protocol":"Post-Training"}. This record does not assert independent reproduction. ### Local threshold balancing (Ours) ANN Accuracy: 69.76. Result ID: 69947df63a6c3148491d9777. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Architecture":"ResNet-18","Evaluation Protocol":"Post-Training"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.