File size: 5,129 Bytes
c989a4a | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 | ---
language:
- sl
license: apache-2.0
library_name: transformers
tags:
- automatic-speech-recognition
- speech
- whisper
- slovenian
- seq2seq
datasets:
- mozilla-foundation/common_voice_17_0
- yuriyvnv/synthetic_asr_et_sl
base_model: openai/whisper-large-v3
pipeline_tag: automatic-speech-recognition
model-index:
- name: whisper-large-v3-slovenian
results:
- task:
type: automatic-speech-recognition
name: Speech Recognition
dataset:
name: Common Voice 17.0 (sl) - Test
type: mozilla-foundation/common_voice_17_0
config: sl
split: test
metrics:
- type: wer
value: 15.65
name: WER (raw)
- type: wer
value: 13.24
name: WER (normalized)
- type: cer
value: 3.88
name: CER (raw)
- task:
type: automatic-speech-recognition
name: Speech Recognition
dataset:
name: FLEURS (sl) - Test
type: google/fleurs
config: sl_si
split: test
metrics:
- type: wer
value: 40.46
name: WER (raw)
- type: wer
value: 24.52
name: WER (normalized)
- type: cer
value: 11.98
name: CER (raw)
---
# Whisper-large-v3 Slovenian
Fine-tuned [OpenAI Whisper-large-v3](https://huggingface.co/openai/whisper-large-v3) for Slovenian automatic speech recognition, augmented with TTS-generated synthetic data.
This model is part of the paper: *"Synthetic Speech Augmentation for Low-Resource Estonian and Slovenian ASR: Comparing Parakeet-TDT and Whisper"* (Interspeech 2026). Paper coming soon.
> **Note:** Fine-tuning Whisper on CommonVoice data degrades FLEURS performance for Slovenian. Zero-shot Whisper (37.02% raw / 19.18% normalized) outperforms all fine-tuned variants on FLEURS. This is because Whisper's pre-trained model handles FLEURS-style punctuation and casing, which fine-tuning on CV-style data erodes. On the in-domain CV17 test set, this model provides substantial improvements over both zero-shot and CV-only fine-tuning.
## Model Description
- **Architecture:** Encoder-decoder Transformer (Whisper)
- **Parameters:** 1.55B
- **Tokenizer:** 51,865-token byte-level BPE
- **Base model:** `openai/whisper-large-v3`
- **Fine-tuning data:** CommonVoice 17.0 Slovenian + ~5,850 synthetic sentences (paraphrase + domain expansion, without morphological category)
- **Training config:** CV + Synth No Morph (synthetic corpus excluding morphological augmentation)
## Evaluation Results
### Raw WER/CER (no text normalization)
| Test Set | WER | CER |
|----------|-----|-----|
| CommonVoice 17 Test | **15.65** | **3.88** |
| CommonVoice 17 Val | **15.08** | **3.60** |
| FLEURS Test | 40.46 | 11.98 |
### Normalized WER/CER (lowercase + punctuation removal)
| Test Set | WER | CER |
|----------|-----|-----|
| CommonVoice 17 Test | **13.24** | **3.48** |
| CommonVoice 17 Val | **12.98** | **3.24** |
| FLEURS Test | 24.52 | 8.32 |
### Improvement over baselines
| Comparison | CV17 Test (WER) | FLEURS Test (WER) |
|-----------|-----------------|-------------------|
| vs. Zero-shot | -5.55 pp | +3.44 pp* |
| vs. CV-only fine-tuning | -3.66 pp | -6.33 pp |
*FLEURS degrades vs. zero-shot due to domain mismatch (see note above), but improves substantially vs. CV-only fine-tuning.
All CV17 improvements are statistically significant (paired bootstrap, p < 0.001, n = 100,000).
## Usage
```python
from transformers import WhisperForConditionalGeneration, WhisperProcessor
import torch
# Load model and processor
model = WhisperForConditionalGeneration.from_pretrained("yuriyvnv/whisper-large-v3-slovenian")
processor = WhisperProcessor.from_pretrained("yuriyvnv/whisper-large-v3-slovenian")
# Load audio (16kHz mono)
import librosa
audio, sr = librosa.load("audio.wav", sr=16000)
# Transcribe
input_features = processor(audio, sampling_rate=16000, return_tensors="pt").input_features
with torch.no_grad():
predicted_ids = model.generate(input_features, language="sl", task="transcribe")
transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True)[0]
print(transcription)
```
## Training Details
- **Optimizer:** AdamW fused (lr=5e-5)
- **Schedule:** Linear decay with 10% warmup
- **Effective batch size:** 128 (64 x 2 gradient accumulation)
- **Epochs:** 5
- **Best model:** selected by eval_loss
- **Precision:** bf16
- **Seed:** 42
## Synthetic Data Augmentation
The synthetic training data was generated using a three-stage pipeline:
1. **Text generation:** GPT-5-mini generates diverse sentences across paraphrase, domain expansion, and morphological categories
2. **LLM-as-judge validation:** Each sentence validated for grammaticality, naturalness, and language purity
3. **Speech synthesis:** OpenAI gpt-4o-mini-tts with 11-voice rotation
Dataset: [yuriyvnv/synthetic_asr_et_sl](https://huggingface.co/datasets/yuriyvnv/synthetic_asr_et_sl)
## Acknowledgments
- Base model: [OpenAI Whisper-large-v3](https://huggingface.co/openai/whisper-large-v3)
- Training data: [Mozilla Common Voice 17.0](https://commonvoice.mozilla.org/)
- Evaluation: [Google FLEURS](https://huggingface.co/datasets/google/fleurs)
|