Hugging Face's logo

chappM
/
whisper-amharic-small-v2

Automatic Speech Recognition
Safetensors
Amharic
whisper
speech
amharic
asr
Eval Results (legacy)
Model card Files
xet
 Community

Whisper Amharic Small v2

Fine-tuned Whisper Small model for Amharic speech recognition using Mozilla Common Voice dataset.

Performance

  • WER: 69% (improved from 77% in v1)
  • Dataset: Mozilla Common Voice Amharic
  • Base Model: openai/whisper-small

Usage 

from transformers import WhisperProcessor, WhisperForConditionalGeneration

processor = WhisperProcessor.from_pretrained("chappM/whisper-amharic-small-v2")
model = WhisperForConditionalGeneration.from_pretrained("chappM/whisper-amharic-small-v2")

Optimized for Amharic speech recognition tasks.

Downloads last month
240
Safetensors
Model size
0.2B params
Tensor type
F32
·
Inference Providers NEW
Automatic Speech Recognition
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for chappM/whisper-amharic-small-v2

Base model

openai/whisper-small
Finetuned
(3503)
this model

Dataset used to train chappM/whisper-amharic-small-v2

Space using chappM/whisper-amharic-small-v2 1

Evaluation results