unsloth-stt
Unsloth supports fine-tuning for Speech-to-Text (STT) models like OpenAI Whisper. By applying its optimized LoRA pipeline to Whisper architecture, Unsloth achieves 1.5x faster training with 50% less memory usage compared to standard methods.
npx skills add majiayu000/claude-skill-registry --skill unsloth-stt --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
--- name: unsloth-stt description: Fine-tuning Speech-to-Text models like Whisper using Unsloth's optimized LoRA pipeline. Triggers: stt, whisper, transcription, audio fine-tuning, speech-to-text, audio normalization. --- ## Overview Unsloth supports fine-tuning for Speech-to-Text (STT) models like OpenAI Whisper. By applying its optimized LoRA pipeline to Whisper architecture, Unsloth achieves 1.5x faster training with 50% less memory usage compared to standard methods. ## When to Use - When you need to capture specialized terminology (medical, legal) that base Whisper misses. - When adapting models to specific accents or dialects. - When fine-tuning large models (like whisper-large-v3) on limited consumer hardware. ## Decision Tree 1. Is transcription accuracy low on zero-shot inference? - Yes: Proceed with STT fine-tuning. 2. Is audio recorded at a non-standard sample rate? - Yes: Resample to 16kHz before training. 3. Using large-v3? - Yes: Load in 4-bit and apply LoRA to cross-attention layers. ## Workflows ### Whisper STT Data Preprocessing 1. Load audio files using `datasets.Audio` feature to handle on-the-fly decoding. 2. Resample all training audio to 16kHz to avoid sample
- Overview
- When to Use
- Decision Tree
- Workflows
- Whisper STT Data Preprocessing
- Fine-tuning Whisper with Unsloth
- Non-Obvious Insights
- Evidence
- Scripts
- Dependencies
- References
What does the unsloth-stt skill do?
Unsloth supports fine-tuning for Speech-to-Text (STT) models like OpenAI Whisper. By applying its optimized LoRA pipeline to Whisper architecture, Unsloth achieves 1.5x faster training with 50% less memory usage compared to standard methods.
How do I install it?
Run `npx skills add majiayu000/claude-skill-registry --skill unsloth-stt --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From majiayu000/claude-skill-registry, a repository with 534 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.