Dataset Viewer (First 5GB)
Auto-converted to Parquet
text
string
0.87884
0.71759
0.86554
0.85888
0.74540
0.87758
END_OF_DOCUMENT
0.86418
0.86418
0.86418
0.87270
0.89489
0.86117
0.81869
0.86924
0.87769
0.86674
END_OF_DOCUMENT
0.62245
0.56956
0.44456
0.83348
0.70374
0.79478
0.87154
0.88212
0.77200
0.62292
0.71669
0.83147
0.79554
0.77286
0.85526
0.64783
0.86646
0.82657
0.80871
0.88877
0.82096
0.52197
0.62109
0.85256
0.84553
0.69529
0.69295
0.76681
0.87207
0.71314
0.83441
0.79380
0.86970
0.83722
0.43179
0.82070
0.55077
0.88222
0.78750
0.84346
0.74871
0.87005
0.77461
0.83815
0.89286
0.87037
0.83225
0.62627
0.85063
0.87282
0.80903
0.79391
0.75159
0.67119
0.88106
0.71495
0.83561
0.80267
0.87764
0.79186
0.88026
0.83479
0.72592
0.44627
0.54781
0.69050
0.83789
0.49007
0.86152
0.76570
0.81362
0.85567
0.62100
0.75373
0.85042
0.62850
0.69120
0.72936
0.74913
0.87367
0.78124
0.86711
End of preview. Expand in Data Studio

This dataset contains quality estimation scores (done with Unbabel/wmt22-cometkiwi-da) for the translations with Opus into Spanish.

The dataset follows the plain text format here: https://github.com/Helsinki-NLP/translate-fineweb/blob/main/maxidl/nemotron-cc-english-run1/translated/txt/spa/README.md

Each line is the QE score for the corresponding line in the translated Spanish data. END_OF_DOCUMENT are inserted to delimit document boundaries (note that END_OF_DOCUMENT markers were translated as FIN DE DOCUMENTO in the Spanish text). Each shard here corresponds to one shard in the translated data.

If document-level scores are desired, these have to be an aggregated version of sentence-level scores.

Downloads last month
106