1fdfaff096
Proper nouns (Córdoba, Cancún) and dropped words came from decoding the committed transcript with the small model. The image already ships large-v3-turbo, so the final decode now uses it with beam_size=5, a temperature fallback ladder, and a proper-noun/accents initial_prompt that fixes first-pass capitalization and diacritics across EN/ES/RU; the rolling previews stay on tiny at greedy so the on-the-fly feel is unchanged. The accurate decode runs in the speculative predecode during the end-of-speech silence and is cache-reused at commit, so perceived latency stays low. All decode knobs are env-overridable for on-device tuning (beam/temperature/prompt), with small as the guaranteed-present rollback if turbo underperforms on the Jetson. 206 STT tests pass. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BvMSXH8VH2tMWXanb8SJdf
titan-iac
Flux-managed Kubernetes desired-state config for bstein.dev.
Canonical source URL:
ssh://git@scm.bstein.dev:2242/atlas/titan-iac.git
Scope
This repo contains cluster configuration consumed by Flux:
- platform/infrastructure manifests
- service manifests and kustomizations
- operational scripts for render/reconcile workflows
Apply model
I use Git + Flux as the source of truth and avoid manual in-cluster edits for durable changes.