# Code-switched text synthesis on Hindi-English (Hi-En) (test)

Source: [Code-switched text synthesis on Hindi-English (Hi-En) (test)](<https://www.sota2.com/research/sota/code-switched-text-synthesis-on-hindi-english-hi-en-test>)

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

Data updated: 2026-02-26T05:38:49.438Z

Evaluation of code-switched text synthesis performance for the Hindi-English (Hi-En) test set, reporting metrics such as CS Rate, F, S, and Geo. Mean.

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 | Code-switched text synthesis |
| Dataset (including any stated split) | Hindi-English (Hi-En) (test) |
| Metric | CS Rate |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 6 of 6 |
| 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

- [CS Rate](<https://www.sota2.com/research/sota/code-switched-text-synthesis-on-hindi-english-hi-en-test.md?metric=CS%20Rate>): Higher is better; unit: PERCENT.

- [F](<https://www.sota2.com/research/sota/code-switched-text-synthesis-on-hindi-english-hi-en-test.md?metric=F>): Higher is better; unit: SCALAR.

- [S](<https://www.sota2.com/research/sota/code-switched-text-synthesis-on-hindi-english-hi-en-test.md?metric=S>): Higher is better; unit: SCALAR.

- [Geometric Mean](<https://www.sota2.com/research/sota/code-switched-text-synthesis-on-hindi-english-hi-en-test.md?metric=Geometric%20Mean>): Higher is better; unit: SCALAR.

## Recorded results

| Method | CS Rate | Result date | Report paper |
| --- | --- | --- | --- |
| Translate, Align, then Swap | 98.6 | Not recorded | 68fa6ff81e209131cea483cd |
| Fine-tuned PMMTM on all language pairs | 98 | Not recorded | 68fa6ff81e209131cea483cd |
| Ground truth | 96 | Not recorded | 68fa6ff81e209131cea483cd |
| GLOSS (prefix + self-training) | 93.3 | Not recorded | 68fa6ff81e209131cea483cd |
| GLOSS (prefix) | 87.3 | Not recorded | 68fa6ff81e209131cea483cd |
| Fine-tuned PMMTM on available language pairs | 4 | Not recorded | 68fa6ff81e209131cea483cd |

### Translate, Align, then Swap

CS Rate: 98.6. Result ID: 699103083a6c3148490502fc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6ff81e209131cea483cd.md>)

Recorded evaluation configuration: {"Type":"Unsup."}. This record does not assert independent reproduction.

### Fine-tuned PMMTM on all language pairs

CS Rate: 98. Result ID: 699103083a6c314849050300.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6ff81e209131cea483cd.md>)

Recorded evaluation configuration: {"Type":"Sup."}. This record does not assert independent reproduction.

### Ground truth

CS Rate: 96. Result ID: 699103083a6c314849050301.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6ff81e209131cea483cd.md>)

Recorded evaluation configuration: {"Type":"Reference"}. This record does not assert independent reproduction.

### GLOSS (prefix + self-training)

CS Rate: 93.3. Result ID: 699103083a6c3148490502ff.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6ff81e209131cea483cd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ff81e209131cea483cd.md>)

[Implementation](<https://www.github.com/bsolomon1124/pycld3>)

Recorded evaluation configuration: {"Type":"Zst.","Backbone":"augment-MMT","Self-training (R)":5}. This record does not assert independent reproduction.

### GLOSS (prefix)

CS Rate: 87.3. Result ID: 699103083a6c3148490502fe.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6ff81e209131cea483cd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ff81e209131cea483cd.md>)

[Implementation](<https://www.github.com/bsolomon1124/pycld3>)

Recorded evaluation configuration: {"Type":"Zst.","Backbone":"augment-MMT"}. This record does not assert independent reproduction.

### Fine-tuned PMMTM on available language pairs

CS Rate: 4. Result ID: 699103083a6c3148490502fd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6ff81e209131cea483cd.md>)

Recorded evaluation configuration: {"Type":"Zst."}. This record does not assert independent reproduction.

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