Teaching a Small Model to Stop Spelling Out Numbers
Dictation pipelines break on numbers, not hard words. This episode digs into inverse text normalization (ITN) — the post-processing step that turns spoken forms into written ones — and why the standard solution isn't a generative LLM at all. We walk through two competing approaches to generating training pairs (LLM teachers vs. rule-based rewrite engines), why NVIDIA's production tagger is a token classifier rather than a seq2seq generator, and what that architectural choice means for hallucination risk. Plus: how real teams handle review (spot-checks and quality gates, not full manual…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.