Downloads · 30 days
40
13% of all-time downloads
CAST-ML/whisper-large-v3-mwa-hy
whisper-large-v3-mwa-hy is a automatic speech recognition model from CAST-ML. Use it when you need speech turned into text. The card lists the license as cc-by-4.0.
This is a Whisper Large v3 model fine-tuned on the Modern Western Armenian Armenian dialect for automatic speech recognition (ASR). It can transcribe audio in the Modern Western Armenian dialect with high accuracy.
Downloads · 30 days
40
13% of all-time downloads
All-time downloads
315
Public
Parameters
1.5B
12.3 GB on disk
Likes
0
Public
Click a slice to open those files.
.pth6.2 GB · 50%
From the Hugging Face model README
This is a Whisper Large v3 model fine-tuned on the Modern Western Armenian Armenian dialect for automatic speech recognition (ASR).
It can transcribe audio in the Modern Western Armenian dialect with high accuracy.
openai/whisper-large-v3torch==2.3.1transformers==4.43.0import torch
from transformers import pipeline
model = torch.load(
"model.pth",
weights_only=False,
)
pipe = pipeline(
"automatic-speech-recognition",
model=model,
tokenizer="openai/whisper-large-v3",
feature_extractor="openai/whisper-large-v3",
device="cuda" if torch.cuda.is_available() else "cpu"
)
audio_paths = [
"example1.wav",
"example2.wav"
]
for audio_path in audio_paths:
result = pipe(audio_path)
print(f"Audio: {audio_paths}")
print(f"Transcription: {result['text']}")
Audio: example1.wav
Original sentence: "շարունակել քանի վատ էր վիճակնիս վատ էր չկրցայ դպրոց"
Model transcription: "շարունակել քանի վատ է վիճակնիս վատ էր չի կրցայ դպրոց"
Audio: example2.wav
Original sentence: "Իմ հետս ոչ չէ եղած բայց աս խնդիրը շատ կը լսեմ շրջապատէս"
Model transcription: "իմ հետս ոչ չէ եղած բայց այս խնդիրը շատ կը լսեմ շրջապատս"