Dataset · seanghay
Khmer audio and transcripts
seanghay/khmer_mpwt_speech
Nearly two hours of voiced Khmer phrases from a speech synthesis mobile app.
What's inside
Contains between 1,000 and 10,000 pairs of Khmer audio and text with a total duration of about 1.93 hours (6,957 seconds). Data was sourced from the Ministry of Public Works and Transport mobile app and manually verified, though errors may still occur. Intended exclusively for research purposes.
What you can do with it
- Training or fine-tuning text-to-speech (TTS) models for the Khmer language
- Testing and adapting automatic speech recognition (ASR) models for a low-resource language
- Tasks
- Text → speech
- Languages
- Khmer
- Size
- 1K–10K rows
- Downloads this month
- 217
- Likes
- 11
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
- KrorngAI/mms-tts-khm-finetunedText → speech 105
- Kassiyakk/Khmer_Whisper_Fine_TuneAutomatic speech recognition 87
- sokhapen/whisper-small-khmer-v6-ct2 41
- PhanithLIM/whisper-base-aug-20-april-lightning-v1Automatic speech recognition 35
- PhanithLIM/whisper-tiny-aug-7-may-lightning-v1Automatic speech recognition 13
How to load it
from datasets import load_dataset
ds = load_dataset("seanghay/khmer_mpwt_speech")