Hugging Face's logo
Yuriy Perezhohin's picture
Building on HF
12 5 30

Yuriy Perezhohin PRO

yuriyvnv
flph123's profile picture webxos's profile picture manassehzw's profile picture
Ā·
https://scholar.google.com/citations?user=I5uzFtwAAAAJ&hl=en

AI & ML interests

Automatic Speech Recognition, Embeddings, Code Generation, Synthetic Data Generation and Filtering

Recent Activity

posted an update 5 days ago
šŸ”Š Four Qwen3-ASR (0.6B and 1.7B) Fine-Tunes for Portuguese and Dutch. Both the 1.7B and 0.6B variants of Alibaba's Qwen3-ASR, fine-tuned for European Portuguese and Dutch and bundled in a single collection. šŸ”— Collection: https://huggingface.co/collections/yuriyvnv/qwen-asr-for-portuguese-and-dutch-17b-and-06b Headline numbers ā€” Common Voice 22 test, with the zero-shot baseline. šŸ‡µšŸ‡¹ Qwen3-ASR-1.7B-PT ā€” 12.91% ā†’ 8.50% WER (-34%) šŸ‡µšŸ‡¹ Qwen3-ASR-0.6B-PT ā€” 18.26% ā†’ 11.85% WER (-35%) šŸ‡³šŸ‡± Qwen3-ASR-1.7B-NL ā€” 6.68% ā†’ 5.28% WER (-21%) šŸ‡³šŸ‡± Qwen3-ASR-0.6B-NL ā€” 12.46% ā†’ 8.31% WER (-33%) The 0.6B variants are the more interesting half of the release. They give up only a few WER points compared to the 1.7B at a third of the parameters ā€” relevant for edge hardware, CPU inference, or anywhere keeping inference cost down. The Dutch 0.6B in particular lands at 8.3% WER on CV22, competitive with much larger systems. The Dutch 1.7B started from a strong 6.7% zero-shot, so the absolute gain is smaller ā€” Qwen already handles Dutch well, and the fine-tune mostly sharpens it on Common Voice's casing and punctuation conventions. Training stuck close to Qwen's official SFT recipe (lr 2e-5, linear schedule, 2% warmup, bf16, gradient checkpointing on a single H100). The data is the differentiator: Common Voice 22 train + validation augmented with synthetic OpenAI-TTS speech, filtered by the WAVe multimodal embedding model that scores clips at the word level and drops the ones that don't align well with their transcripts. šŸ“¦ Full pipeline ā€” synthetic data generation, WAVe filtering, training scripts, evaluation protocol ā€” is open-source: github.com/yuriyvnv/TTS-Augmented-ASR @hf-audio . #asr #speech #parakeet #nvidia #nemo #multilingual #fine-tuning #commonvoice
updated a model 5 days ago
yuriyvnv/Qwen3-ASR-1.7B-NL
updated a collection 5 days ago
Qwen-ASR for Portuguese & Dutch, 1.7B & 0.6B
View all activity

Organizations

Blog-explorers's profile picture