smol-audio: A Colab-Friendly Notebook Collection for Fine-Tuning Whisper, Parakeet, Voxtral, Granite Speech, and Audio Flamingo 3

Deep-unlearning released smol-audio, an Apache-2.0 collection of self-contained Jupyter notebooks for audio AI work.

The notebooks cover fine-tuning ASR models, LoRA, CTC, prompt masking, and practical workflows for speech and audio tasks.

They reproduce and adapt models including Whisper, Parakeet, Voxtral, Granite Speech, and Audio Flamingo 3.

Designed to run in Colab, smol-audio makes modern audio model adaptation more transparent and accessible.