# Semi-supervised Image Classification on ImageNet (10% labeled) Source: [Semi-supervised Image Classification on ImageNet (10% labeled)]() Collection index: [llms.txt]() Data updated: 2026-02-26T06:07:06.425Z Evaluation of Semi-supervised Image Classification performance using the 10% labeled subset of ImageNet, measured by Top-1 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 | Semi-supervised Image Classification | | Dataset (including any stated split) | ImageNet (10% labeled) | | Metric | Top-1 Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 35 of 35 | | 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](): Higher is better; unit: PERCENT. - [Top-5 Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | Top-1 Accuracy | Result date | Report paper | | --- | --- | --- | --- | | PAWS | 79 | Not recorded | 68fa6a721e209131cea282ca | | PAWS | 78.5 | Not recorded | 68fa6a721e209131cea282ca | | PAWS | 77.8 | Not recorded | 68fa6a721e209131cea282ca | | PAWS | 77 | Not recorded | 68fa6a721e209131cea282ca | | BYOL | 75.7 | Not recorded | 68fa6a721e209131cea282ca | | SimCLR | 74.4 | Not recorded | 68fa6a721e209131cea282ca | | SimCLRv2 | 73.9 | Not recorded | 68fa6a721e209131cea282ca | | BYOL | 73.5 | Not recorded | 68fa6a721e209131cea282ca | | FNC | 71.1 | Not recorded | 68fa6b871e209131cea2ee36 | | SwAV | 70.2 | Not recorded | 68fa6b871e209131cea2ee36 | | SwAV | 70.2 | Not recorded | 68fa77061e209131cea70946 | | BT | 69.7 | Not recorded | 68fa77061e209131cea70946 | | BYOL | 68.8 | Not recorded | 68fa6b871e209131cea2ee36 | | BYOL | 68.8 | Not recorded | 68fa77061e209131cea70946 | | SimCLR v2 | 68.4 | Not recorded | 68fa6b871e209131cea2ee36 | | SynCo | 66.6 | Not recorded | 68fa77061e209131cea70946 | | MoCo-v2 (repr.) | 66.1 | Not recorded | 68fa77061e209131cea70946 | | MOCHI (repr.) | 65.7 | Not recorded | 68fa77061e209131cea70946 | | SimCLR v1 | 65.6 | Not recorded | 68fa6b871e209131cea2ee36 | | DATA (7G~8G) | 65.6 | Not recorded | 68fa6c1b1e209131cea32412 | | SimCLR | 65.6 | Not recorded | 68fa77061e209131cea70946 | | DATA (6G~7G) | 65.4 | Not recorded | 68fa6c1b1e209131cea32412 | | DATA (5G~6G) | 65.1 | Not recorded | 68fa6c1b1e209131cea32412 | | DATA (4G~5G) | 64.3 | Not recorded | 68fa6c1b1e209131cea32412 | | DATA (3G~4G) | 63.4 | Not recorded | 68fa6c1b1e209131cea32412 | | DATA (2G~3G) | 63 | Not recorded | 68fa6c1b1e209131cea32412 | | DATA (1G~2G) | 61.9 | Not recorded | 68fa6c1b1e209131cea32412 | | ResNet-50 | 61.8 | Not recorded | 68fa6c1b1e209131cea32412 | | PIRL | 60.4 | Not recorded | 68fa77061e209131cea70946 | | PIRL | 57.2 | Not recorded | 68fa6b871e209131cea2ee36 | | Supervised | 56.4 | Not recorded | 68fa77061e209131cea70946 | | Supervised | 48.4 | Not recorded | 68fa6b871e209131cea2ee36 | ### PAWS Top-1 Accuracy: 79. Result ID: 698e353dffc3901de09bab42. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Architecture":"ResNet-50 (4x)","Epochs":"200"}. This record does not assert independent reproduction. ### PAWS Top-1 Accuracy: 78.5. Result ID: 698e353dffc3901de09bab41. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Architecture":"ResNet-50 (4x)","Epochs":"100"}. This record does not assert independent reproduction. ### PAWS Top-1 Accuracy: 77.8. Result ID: 698e353dffc3901de09bab3e. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Architecture":"ResNet-50 (2x)","Epochs":"200"}. This record does not assert independent reproduction. ### PAWS Top-1 Accuracy: 77. Result ID: 698e353dffc3901de09bab3d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Architecture":"ResNet-50 (2x)","Epochs":"100"}. This record does not assert independent reproduction. ### BYOL Top-1 Accuracy: 75.7. Result ID: 698e353dffc3901de09bab40. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Architecture":"ResNet-50 (4x)","Epochs":"1000"}. This record does not assert independent reproduction. ### SimCLR Top-1 Accuracy: 74.4. Result ID: 698e353dffc3901de09bab3f. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Architecture":"ResNet-50 (4x)","Epochs":"1000"}. This record does not assert independent reproduction. ### SimCLRv2 Top-1 Accuracy: 73.9. Result ID: 698e353dffc3901de09bab3c. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Architecture":"ResNet-50 (2x)","Epochs":"800"}. This record does not assert independent reproduction. ### BYOL Top-1 Accuracy: 73.5. Result ID: 698e353dffc3901de09bab3b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Architecture":"ResNet-50 (2x)","Epochs":"1000"}. This record does not assert independent reproduction. ### FNC Top-1 Accuracy: 71.1. Result ID: 698e94d8ffc3901de09ed7a5. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"ResNet-50","Learning Type":"Representation Learning / Contrastive"}. This record does not assert independent reproduction. ### SwAV Top-1 Accuracy: 70.2. Result ID: 698e94d8ffc3901de09ed7a7. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"ResNet-50"}. This record does not assert independent reproduction. ### SwAV Top-1 Accuracy: 70.2. Result ID: 699471fe3a6c3148491d499b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Epochs":800,"Backbone":"ResNet-50"}. This record does not assert independent reproduction. ### BT Top-1 Accuracy: 69.7. Result ID: 699471fe3a6c3148491d4999. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Epochs":1000,"Backbone":"ResNet-50"}. This record does not assert independent reproduction. ### BYOL Top-1 Accuracy: 68.8. Result ID: 698e94d8ffc3901de09ed7a6. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"ResNet-50"}. This record does not assert independent reproduction. ### BYOL Top-1 Accuracy: 68.8. Result ID: 699471fe3a6c3148491d499a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Epochs":1000,"Backbone":"ResNet-50"}. This record does not assert independent reproduction. ### SimCLR v2 Top-1 Accuracy: 68.4. Result ID: 698e94d8ffc3901de09ed7a4. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"ResNet-50","Learning Type":"Representation Learning / Contrastive"}. This record does not assert independent reproduction. ### SynCo Top-1 Accuracy: 66.6. Result ID: 699471fe3a6c3148491d499e. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Epochs":800,"Backbone":"ResNet-50"}. This record does not assert independent reproduction. ### MoCo-v2 (repr.) Top-1 Accuracy: 66.1. Result ID: 699471fe3a6c3148491d499c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Epochs":800,"Backbone":"ResNet-50"}. This record does not assert independent reproduction. ### MOCHI (repr.) Top-1 Accuracy: 65.7. Result ID: 699471fe3a6c3148491d499d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Epochs":800,"Backbone":"ResNet-50"}. This record does not assert independent reproduction. ### SimCLR v1 Top-1 Accuracy: 65.6. Result ID: 698e94d8ffc3901de09ed7a3. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"ResNet-50","Learning Type":"Representation Learning / Contrastive"}. This record does not assert independent reproduction. ### DATA (7G~8G) Top-1 Accuracy: 65.6. Result ID: 698ed988ffc3901de0a21a39. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Params":"45.2M","Fine-tuning epochs":"20"}. This record does not assert independent reproduction. ### SimCLR Top-1 Accuracy: 65.6. Result ID: 699471fe3a6c3148491d4998. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Epochs":1000,"Backbone":"ResNet-50"}. This record does not assert independent reproduction. ### DATA (6G~7G) Top-1 Accuracy: 65.4. Result ID: 698ed988ffc3901de0a21a38. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Params":"45.4M","Fine-tuning epochs":"20"}. This record does not assert independent reproduction. ### DATA (5G~6G) Top-1 Accuracy: 65.1. Result ID: 698ed988ffc3901de0a21a37. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Params":"43.4M","Fine-tuning epochs":"20"}. This record does not assert independent reproduction. ### DATA (4G~5G) Top-1 Accuracy: 64.3. Result ID: 698ed988ffc3901de0a21a36. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Params":"37.0M","Fine-tuning epochs":"20"}. This record does not assert independent reproduction. ### DATA (3G~4G) Top-1 Accuracy: 63.4. Result ID: 698ed988ffc3901de0a21a35. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Params":"33.5M","Fine-tuning epochs":"20"}. This record does not assert independent reproduction. ### DATA (2G~3G) Top-1 Accuracy: 63. Result ID: 698ed988ffc3901de0a21a34. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Params":"19.5M","Fine-tuning epochs":"20"}. This record does not assert independent reproduction. ### DATA (1G~2G) Top-1 Accuracy: 61.9. Result ID: 698ed988ffc3901de0a21a33. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Params":"14.7M","Fine-tuning epochs":"20"}. This record does not assert independent reproduction. ### ResNet-50 Top-1 Accuracy: 61.8. Result ID: 698ed988ffc3901de0a21a32. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Params":"25.5M","Pre-training":"MoCo v2","Fine-tuning epochs":"20"}. This record does not assert independent reproduction. ### PIRL Top-1 Accuracy: 60.4. Result ID: 699471fe3a6c3148491d4997. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Epochs":800,"Backbone":"ResNet-50"}. This record does not assert independent reproduction. ### PIRL Top-1 Accuracy: 57.2. Result ID: 698e94d8ffc3901de09ed7a1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"ResNet-50","Learning Type":"Representation Learning / Contrastive"}. This record does not assert independent reproduction. ### Supervised Top-1 Accuracy: 56.4. Result ID: 699471fe3a6c3148491d4996. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"ResNet-50"}. This record does not assert independent reproduction. ### Supervised Top-1 Accuracy: 48.4. Result ID: 698e94d8ffc3901de09ed79e. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"ResNet-50"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.