CC BY 4.0 synthetic dataset
Пограничные случаи субтитров
Синтетические пограничные случаи тайминга и текста.
Итог проверки
{
"records": 24,
"train": 18,
"evaluation": 6,
"synthetic_only": true,
"leaked_groups": 0
}Примеры записей
[
{
"record_id": "subtitle-edge-cases-01-1",
"group_id": "subtitle-edge-cases-group-01",
"split": "evaluation",
"label_en": "overlap positive example",
"label_ru": "overlap: корректный пример",
"synthetic": true,
"payload": {
"case": "overlap",
"start_ms": 0,
"end_ms": 800,
"expected_valid": true
}
},
{
"record_id": "subtitle-edge-cases-01-2",
"group_id": "subtitle-edge-cases-group-01",
"split": "evaluation",
"label_en": "overlap negative example",
"label_ru": "overlap: ошибочный пример",
"synthetic": true,
"payload": {
"case": "overlap",
"start_ms": 0,
"end_ms": -1,
"expected_valid": false
}
},
{
"record_id": "subtitle-edge-cases-02-1",
"group_id": "subtitle-edge-cases-group-02",
"split": "train",
"label_en": "zero duration positive example",
"label_ru": "zero duration: корректный пример",
"synthetic": true,
"payload": {
"case": "zero duration",
"start_ms": 1000,
"end_ms": 1800,
"expected_valid": true
}
},
{
"record_id": "subtitle-edge-cases-02-2",
"group_id": "subtitle-edge-cases-group-02",
"split": "train",
"label_en": "zero duration negative example",
"label_ru": "zero duration: ошибочный пример",
"synthetic": true,
"payload": {
"case": "zero duration",
"start_ms": 1000,
"end_ms": 999,
"expected_valid": false
}
}
]