Hmong Screener — XLS-R 300M

XLS-R 300M fine-tuned for Hmong CTC speech recognition.

from transformers import AutoModelForCTC, AutoProcessor

processor = AutoProcessor.from_pretrained("twangodev/hmong-screener")
model = AutoModelForCTC.from_pretrained("twangodev/hmong-screener").eval()

Use mono audio at 16 kHz. Word and tone screening requires the upstream scoring code and assets.

Downloads last month: 3

Safetensors

Model size

0.3B params

Tensor type

F32

Model tree for twangodev/hmong-screener

Base model

facebook/wav2vec2-xls-r-300m

Finetuned

(921)

this model

Paper for twangodev/hmong-screener

SITA: Learning Speaker-Invariant and Tone-Aware Speech Representations for Low-Resource Tonal Languages

Paper • 2601.09050 • Published Jan 14