google/WaxalNLP
Viewer • Updated • 2.56M • 44.3k • 227
Fine-tuned openai/whisper-medium for speech recognition on English and Nigerian Pidgin (Naijá).
from transformers import pipeline
pipe = pipeline(
"automatic-speech-recognition",
model="Tamynator/whisper-medium-en-pcm",
chunk_length_s=30,
)
result = pipe("audio.wav")
print(result["text"])
Training Data
Custom English + Pidgin recordings
AfriSpeech
WAXAL Pidgin TTS