Low-resource ASR · Whisper v3-turbo + LoRA
Ng'alal
Kalenjin.
A LoRA fine-tune of Whisper large-v3-turbo. Speak into the mic, upload a clip, or hear a sample — get a transcript in seconds.
26.4%Char error rate
~3sWarm latency
60sClip limit
Console — IdleL4 GPU · Modal
Press to speak
up to 60s