disco-eth/WorldSpeech
Viewer • Updated • 16.5M • 23.9k • 28
🏗️ Building on HF
manasseh
manassehzw
AI & ML interests
Voice Models, TTS, Conversational AI, GPT
Recent Activity
liked a dataset about 14 hours ago
disco-eth/WorldSpeech liked a dataset about 14 hours ago
African-Languages-Lab/all-lab-speech liked a dataset about 23 hours ago
AfriSpeech/open-bible-speech-africanOrganizations
None yet
liked 2
datasets about 14 hours ago
liked 2
datasets about 23 hours ago
liked a
model 19 days ago
liked a
model 26 days ago
liked a
model 27 days ago
liked 3
models about 2 months ago
updated a
model 3 months ago
published a
model 3 months ago
updated a
model 3 months ago
updated a
dataset 3 months ago
published a
dataset 3 months ago
updated a
model 3 months ago
published a
model 3 months ago
Thank you! this is a huge help, let me play around with the repo and see what results i can get. I'll definitely reach out if i need any other assistance.
reacted to yuriyvnv's post with 🚀 4 months ago
Post
640
🎙️Parakeet-TDT Fine Tuning: 4 New ASR Models
Four fine-tuned versions of NVIDIA's Parakeet-TDT-0.6B-v3 for Dutch, Portuguese, Estonian, and Slovenian — among the first community fine-tunes of this architecture for the aforementioned languages
📊 Results on Common Voice 17 test sets:
🇸🇮 Slovenian: 50.49% → 11.56% WER (-77%)
🇵🇹 Portuguese: 15.86% → 10.71% WER (-32%)
🇪🇪 Estonian: 27.15% → 21.03% WER (-23%)
🇳🇱 Dutch: 5.99% → 5.33% WER (-11%)
All models output cased text with punctuation.
🔗 Models:
🇳🇱 yuriyvnv/parakeet-tdt-0.6b-dutch
🇵🇹 yuriyvnv/parakeet-tdt-0.6b-portuguese
🇪🇪 yuriyvnv/parakeet-tdt-0.6b-estonian
🇸🇮 yuriyvnv/parakeet-tdt-0.6b-slovenian
🏗️ Training: Common Voice 17 + synthetic speech (OpenAI TTS), filtered with WAVe (yuriyvnv/WAVe-1B-Multimodal-PT) for quality. AdamW + cosine annealing, bf16-mixed precision, early stopping on val WER. Timestamps and long-form audio supported.
@hf-audio @NVIDIADev
#asr #speech #parakeet #nvidia #nemo #multilingual #fine-tuning #commonvoice
Four fine-tuned versions of NVIDIA's Parakeet-TDT-0.6B-v3 for Dutch, Portuguese, Estonian, and Slovenian — among the first community fine-tunes of this architecture for the aforementioned languages
📊 Results on Common Voice 17 test sets:
🇸🇮 Slovenian: 50.49% → 11.56% WER (-77%)
🇵🇹 Portuguese: 15.86% → 10.71% WER (-32%)
🇪🇪 Estonian: 27.15% → 21.03% WER (-23%)
🇳🇱 Dutch: 5.99% → 5.33% WER (-11%)
All models output cased text with punctuation.
import nemo.collections.asr as nemo_asr
model = nemo_asr.models.ASRModel.from_pretrained(
"yuriyvnv/parakeet-tdt-0.6b-dutch"
)
output = model.transcribe(["audio.wav"])
print(output[0].text)🔗 Models:
🇳🇱 yuriyvnv/parakeet-tdt-0.6b-dutch
🇵🇹 yuriyvnv/parakeet-tdt-0.6b-portuguese
🇪🇪 yuriyvnv/parakeet-tdt-0.6b-estonian
🇸🇮 yuriyvnv/parakeet-tdt-0.6b-slovenian
🏗️ Training: Common Voice 17 + synthetic speech (OpenAI TTS), filtered with WAVe (yuriyvnv/WAVe-1B-Multimodal-PT) for quality. AdamW + cosine annealing, bf16-mixed precision, early stopping on val WER. Timestamps and long-form audio supported.
@hf-audio @NVIDIADev
#asr #speech #parakeet #nvidia #nemo #multilingual #fine-tuning #commonvoice
- 3 replies
Amazing, impressive work!
I’m interested in fine-tuning Parakeet for low-resource African languages. could not find any NVIDIA fine-tuning guide yet, would you be open to sharing any pointers on your training setup, workflow, or resources you found helpful?