# News Document Classification on MLDoc (test)

Source: [News Document Classification on MLDoc (test)](<https://www.sota2.com/research/sota/news-document-classification-on-mldoc-test>)

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

Data updated: 2026-02-26T05:42:22.542Z

Evaluation of news document classification performance across multiple languages (French, German, English, Spanish, Chinese, Russian) on the test split of the MLDoc dataset.

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 | News document classification |
| Dataset (including any stated split) | MLDoc (test) |
| Metric | Error Rate (FR) |
| Unit | PERCENT |
| Direction | Lower 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

- [Error Rate (FR)](<https://www.sota2.com/research/sota/news-document-classification-on-mldoc-test.md?metric=Error%20Rate%20(FR)>): Lower is better; unit: PERCENT.

- [Error Rate (DE)](<https://www.sota2.com/research/sota/news-document-classification-on-mldoc-test.md?metric=Error%20Rate%20(DE)>): Lower is better; unit: PERCENT.

- [Error Rate (EN)](<https://www.sota2.com/research/sota/news-document-classification-on-mldoc-test.md?metric=Error%20Rate%20(EN)>): Lower is better; unit: PERCENT.

- [Error Rate (ES)](<https://www.sota2.com/research/sota/news-document-classification-on-mldoc-test.md?metric=Error%20Rate%20(ES)>): Lower is better; unit: PERCENT.

- [Error Rate (ZH)](<https://www.sota2.com/research/sota/news-document-classification-on-mldoc-test.md?metric=Error%20Rate%20(ZH)>): Lower is better; unit: PERCENT.

- [Error Rate (RU)](<https://www.sota2.com/research/sota/news-document-classification-on-mldoc-test.md?metric=Error%20Rate%20(RU)>): Lower is better; unit: PERCENT.

## Recorded results

| Method | Error Rate (FR) | Result date | Report paper |
| --- | --- | --- | --- |
| XLM\_ft Monolingual Baseline (UDA) | 3.95 | Not recorded | 68fa66661e209131cea10217 |
| XLM\_ft (UDA+Self) | 4.6 | Not recorded | 68fa66661e209131cea10217 |
| XLM\_ft (UDA) | 4.65 | Not recorded | 68fa66661e209131cea10217 |
| XLM (UDA+Self) | 4.75 | Not recorded | 68fa66661e209131cea10217 |
| XLM (UDA) | 4.97 | Not recorded | 68fa66661e209131cea10217 |
| XLM Monolingual Baseline (Ft) | 5.35 | Not recorded | 68fa66661e209131cea10217 |
| XLM\_ft (Ft) | 6.3 | Not recorded | 68fa66661e209131cea10217 |
| XLM (Ft) | 7.95 | Not recorded | 68fa66661e209131cea10217 |
| pre-XLM sota | 21.97 | Not recorded | 68fa66661e209131cea10217 |

### XLM\_ft Monolingual Baseline (UDA)

Error Rate (FR): 3.95. Result ID: 698cefd48b517eb4331dd234.

[Reporting paper](<https://www.sota2.com/research/paper/68fa66661e209131cea10217.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66661e209131cea10217.md>)

[Implementation](<https://github.com/laiguokun/xlu-data>)

Recorded evaluation configuration: {"Base Model":"XLM\_ft","Train Protocol":"UDA","Mode":"Monolingual","Unlabeled data":true}. This record does not assert independent reproduction.

### XLM\_ft (UDA+Self)

Error Rate (FR): 4.6. Result ID: 698cefd48b517eb4331dd232.

[Reporting paper](<https://www.sota2.com/research/paper/68fa66661e209131cea10217.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66661e209131cea10217.md>)

[Implementation](<https://github.com/laiguokun/xlu-data>)

Recorded evaluation configuration: {"Base Model":"XLM\_ft","Train Protocol":"UDA+Self","Unlabeled data":true}. This record does not assert independent reproduction.

### XLM\_ft (UDA)

Error Rate (FR): 4.65. Result ID: 698cefd48b517eb4331dd231.

[Reporting paper](<https://www.sota2.com/research/paper/68fa66661e209131cea10217.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66661e209131cea10217.md>)

[Implementation](<https://github.com/laiguokun/xlu-data>)

Recorded evaluation configuration: {"Base Model":"XLM\_ft","Train Protocol":"UDA","Unlabeled data":true}. This record does not assert independent reproduction.

### XLM (UDA+Self)

Error Rate (FR): 4.75. Result ID: 698cefd48b517eb4331dd22f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa66661e209131cea10217.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66661e209131cea10217.md>)

[Implementation](<https://github.com/laiguokun/xlu-data>)

Recorded evaluation configuration: {"Base Model":"XLM","Train Protocol":"UDA+Self","Unlabeled data":true}. This record does not assert independent reproduction.

### XLM (UDA)

Error Rate (FR): 4.97. Result ID: 698cefd48b517eb4331dd22e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa66661e209131cea10217.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68a61e209131cea1d69a.md>)

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

Recorded evaluation configuration: {"Base Model":"XLM","Train Protocol":"UDA","Unlabeled data":true}. This record does not assert independent reproduction.

### XLM Monolingual Baseline (Ft)

Error Rate (FR): 5.35. Result ID: 698cefd48b517eb4331dd233.

[Reporting paper](<https://www.sota2.com/research/paper/68fa66661e209131cea10217.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa65491e209131cea0943c.md>)

Recorded evaluation configuration: {"Base Model":"XLM","Train Protocol":"Ft","Mode":"Monolingual","Unlabeled data":false}. This record does not assert independent reproduction.

### XLM\_ft (Ft)

Error Rate (FR): 6.3. Result ID: 698cefd48b517eb4331dd230.

[Reporting paper](<https://www.sota2.com/research/paper/68fa66661e209131cea10217.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66661e209131cea10217.md>)

[Implementation](<https://github.com/laiguokun/xlu-data>)

Recorded evaluation configuration: {"Base Model":"XLM\_ft","Train Protocol":"Ft","Unlabeled data":true}. This record does not assert independent reproduction.

### XLM (Ft)

Error Rate (FR): 7.95. Result ID: 698cefd48b517eb4331dd22d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa66661e209131cea10217.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa65491e209131cea0943c.md>)

Recorded evaluation configuration: {"Base Model":"XLM","Train Protocol":"Ft","Unlabeled data":false}. This record does not assert independent reproduction.

### pre-XLM sota

Error Rate (FR): 21.97. Result ID: 698cefd48b517eb4331dd22c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa66661e209131cea10217.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b7a1e209131cea2e687.md>)

[Implementation](<https://github.com/facebookresearch/LASER>)

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

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