Dataset · seanghay
Aligned speech corpus for the Khmer language
seanghay/khmer_grkpp_speech
A small audio-text aligned dataset in the low-resource Khmer language for research tasks.
What's inside
The dataset contains audio recordings and texts in Khmer (km) that have undergone forced alignment. The total audio duration is 4,446 hours, and the sample size is less than 1,000 records. The author does not own the original data and published the compilation exclusively for low-resource language research.
What you can do with it
- Training and fine-tuning automatic speech recognition (ASR) models for the Khmer language
- Testing audio-text forced alignment systems
- Languages
- Khmer
- Size
- up to 1K rows
- Downloads this month
- 202
- Likes
- 9
Models trained on it
- Thareah/Tha-ASRAutomatic speech recognition 2.5K
- PhanithLIM/whisper-small-aug-28-april-lightning-v1Automatic speech recognition 171
- PhanithLIM/whisper-large-v3-turbo-e3 118
- PhanithLIM/whisper-base-aug-20-april-lightning-v1Automatic speech recognition 35
- Vira21/whisper-tiny-khmerAutomatic speech recognition 13
- PhanithLIM/whisper-tiny-aug-7-may-lightning-v1Automatic speech recognition 13
- sokhapen/whisper-small-khmer-v3-ct2Automatic speech recognition 10
- sokhapen/whisper-small-khmerAutomatic speech recognition 9
How to load it
from datasets import load_dataset
ds = load_dataset("seanghay/khmer_grkpp_speech")