# Unsupervised Disentanglement on CMNIST (test)

Source: [Unsupervised Disentanglement on CMNIST (test)](<https://www.sota2.com/research/sota/unsupervised-disentanglement-on-cmnist-test>)

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

Data updated: 2026-02-26T05:40:12.952Z

Evaluation of unsupervised disentanglement performance on the test split of the CMNIST dataset, using metrics such as DCI, MOD, EXP, LR, and GBT.

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 | Unsupervised Disentanglement |
| Dataset (including any stated split) | CMNIST (test) |
| Metric | DCI |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 7 of 7 |
| 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

- [DCI](<https://www.sota2.com/research/sota/unsupervised-disentanglement-on-cmnist-test.md?metric=DCI>): Higher is better; unit: SCALAR.

- [MOD](<https://www.sota2.com/research/sota/unsupervised-disentanglement-on-cmnist-test.md?metric=MOD>): Higher is better; unit: SCALAR.

- [EXP](<https://www.sota2.com/research/sota/unsupervised-disentanglement-on-cmnist-test.md?metric=EXP>): Higher is better; unit: SCALAR.

- [LR](<https://www.sota2.com/research/sota/unsupervised-disentanglement-on-cmnist-test.md?metric=LR>): Higher is better; unit: SCALAR.

- [GBT](<https://www.sota2.com/research/sota/unsupervised-disentanglement-on-cmnist-test.md?metric=GBT>): Higher is better; unit: SCALAR.

- [Average Score](<https://www.sota2.com/research/sota/unsupervised-disentanglement-on-cmnist-test.md?metric=Average%20Score>): Higher is better; unit: SCALAR.

## Recorded results

| Method | DCI | Result date | Report paper |
| --- | --- | --- | --- |
| VAE | 0.948 | Not recorded | 68fa6b2c1e209131cea2c8e9 |
| β-VAE | 0.945 | Not recorded | 68fa6b2c1e209131cea2c8e9 |
| IP-IRM | 0.917 | Not recorded | 68fa6b2c1e209131cea2c8e9 |
| Factor-VAE | 0.916 | Not recorded | 68fa6b2c1e209131cea2c8e9 |
| β-TCVAE | 0.914 | Not recorded | 68fa6b2c1e209131cea2c8e9 |
| β-AnnealVAE | 0.911 | Not recorded | 68fa6b2c1e209131cea2c8e9 |
| SimCLR | 0.882 | Not recorded | 68fa6b2c1e209131cea2c8e9 |

### VAE

DCI: 0.948. Result ID: 698e6777ffc3901de09d4d88.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2c1e209131cea2c8e9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e2d1e209131cea3e5fb.md>)

Recorded evaluation configuration: {"Mode":"Unsupervised"}. This record does not assert independent reproduction.

### β-VAE

DCI: 0.945. Result ID: 698e6777ffc3901de09d4d89.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2c1e209131cea2c8e9.md>)

Recorded evaluation configuration: {"Mode":"Unsupervised"}. This record does not assert independent reproduction.

### IP-IRM

DCI: 0.917. Result ID: 698e6777ffc3901de09d4d8e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2c1e209131cea2c8e9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b2c1e209131cea2c8e9.md>)

[Implementation](<https://github.com/Wangt-CN/IP-IRM>)

Recorded evaluation configuration: {"Mode":"Unsupervised"}. This record does not assert independent reproduction.

### Factor-VAE

DCI: 0.916. Result ID: 698e6777ffc3901de09d4d8c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2c1e209131cea2c8e9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa65ff1e209131cea0ddd7.md>)

Recorded evaluation configuration: {"Mode":"Unsupervised"}. This record does not assert independent reproduction.

### β-TCVAE

DCI: 0.914. Result ID: 698e6777ffc3901de09d4d8b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2c1e209131cea2c8e9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa65a91e209131cea0bb81.md>)

Recorded evaluation configuration: {"Mode":"Unsupervised"}. This record does not assert independent reproduction.

### β-AnnealVAE

DCI: 0.911. Result ID: 698e6777ffc3901de09d4d8a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2c1e209131cea2c8e9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64441e209131cea039d1.md>)

Recorded evaluation configuration: {"Mode":"Unsupervised"}. This record does not assert independent reproduction.

### SimCLR

DCI: 0.882. Result ID: 698e6777ffc3901de09d4d8d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2c1e209131cea2c8e9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67d71e209131cea18de5.md>)

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

Recorded evaluation configuration: {"Mode":"Standard SSL"}. This record does not assert independent reproduction.

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