metadata
license: mit
task_categories:
- automatic-speech-recognition
language:
- km
Khmer ASR Phase 1 Cache
This repository contains cached training artifacts used for Khmer ASR backbone evaluation.
Source Datasets
- seanghay/khmer_grkpp_speech
- seanghay/km-speech-corpus
- KrorngAI/fleurs_openslr42_mpwt
Contents
- combined_dataset.pt
Usage
import torch
data = torch.load("combined_dataset.pt")