# Clustering on ImageNet-1K 1.28M images (train)

Source: [Clustering on ImageNet-1K 1.28M images (train)](<https://www.sota2.com/research/sota/clustering-on-imagenet-1k-1-28m-images-train>)

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

Data updated: 2026-02-26T05:53:22.628Z

Evaluation of image clustering performance on the training split of the ImageNet-1K dataset, measured using Normalized Mutual Information (NMI), Clustering Accuracy (ACC), and Adjusted Rand Index (ARI).

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 | Clustering |
| Dataset (including any stated split) | ImageNet-1K 1.28M images (train) |
| Metric | NMI |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 3 of 3 |
| 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

- [NMI](<https://www.sota2.com/research/sota/clustering-on-imagenet-1k-1-28m-images-train.md?metric=NMI>): Higher is better; unit: PERCENT.

- [ACC](<https://www.sota2.com/research/sota/clustering-on-imagenet-1k-1-28m-images-train.md?metric=ACC>): Higher is better; unit: PERCENT.

- [ARI](<https://www.sota2.com/research/sota/clustering-on-imagenet-1k-1-28m-images-train.md?metric=ARI>): Higher is better; unit: PERCENT.

## Recorded results

| Method | NMI | Result date | Report paper |
| --- | --- | --- | --- |
| TCL | 67.11 | Not recorded | 68fa6dbf1e209131cea3bd21 |
| TCL | 63.32 | Not recorded | 68fa6dbf1e209131cea3bd21 |
| MoCo (inherited model) | 61.86 | Not recorded | 68fa6dbf1e209131cea3bd21 |

### TCL

NMI: 67.11. Result ID: 698fd97c3a6c314849fc4efd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6dbf1e209131cea3bd21.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6dbf1e209131cea3bd21.md>)

Recorded evaluation configuration: {"Boosting stage":true}. This record does not assert independent reproduction.

### TCL

NMI: 63.32. Result ID: 698fd97c3a6c314849fc4efc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6dbf1e209131cea3bd21.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6dbf1e209131cea3bd21.md>)

Recorded evaluation configuration: {"Boosting stage":false}. This record does not assert independent reproduction.

### MoCo (inherited model)

NMI: 61.86. Result ID: 698fd97c3a6c314849fc4efb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6dbf1e209131cea3bd21.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67371e209131cea14e30.md>)

Recorded evaluation configuration: {"Algorithm":"k-means on extracted features"}. This record does not assert independent reproduction.

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