AdamPodoxin / llm-lora-chapterization Star 0 Code Issues Pull requests Podcast chapterization with LoRA-fine-tuned Qwen3-4B-Instruct: transcripts → hierarchical table of contents. End-to-end pipeline with Whisper ASR, ToC-LLM distillation, two-stage duration-specific fine-tuning, and WindowDiff/ROUGE-L/SBERT evaluation on YTSEG. pytorch lora datasets distillation huggingface chapterization for-portfolio whisper-ai qwen3 ytseg Updated Jun 17, 2026 Python