vh
55602b7251
catalog(dia2): expose full /tts control surface + stable predefined voices
Repoint both dia2 entries from /v1/audio/speech to the wrapper's richer /tts
endpoint (CustomTTSRequest), exposing the levers that fix the random-voice
problem: voice_mode, clone_reference_filename, cfg_scale, temperature, top_p,
cfg_filter_top_k, speed_factor, seed, split_text, chunk_size, transcript,
max_tokens. All defaults are the wrapper's Pydantic blessed values (cfg 3.0 /
temp 1.3 / top_p 0.95 / top_k 35 / speed_factor 0.94 / chunk 300). Fields
grouped (basic/sampling/advanced). dia2 -> version 2 (field-shape change).
Voice stability: Dia2 samples a random speaker per call unless anchored. The
43 curated voices baked at /app/voices aren't reachable from /tts's clone path
(reference_audio dir only), so they're staged into reference_audio; the
clone_reference_filename picker now sources /get_reference_files. voice_mode=
clone + a reference filename pins voice/gender. Verified /tts clone end-to-end
(HTTP 200, Ogg/Opus 24 kHz). README documents the staging + two-instance shape.
2026-05-31 15:18:53 -07:00
..
2026-05-11 15:48:33 -07:00
2026-05-17 08:36:46 -07:00
2026-05-12 13:24:07 -07:00
2026-04-20 22:14:00 -07:00
2026-05-23 18:13:59 -07:00
2026-04-20 22:28:30 -07:00
2026-04-27 21:54:13 -07:00
2026-04-23 22:56:21 -07:00
2026-04-24 16:45:19 -07:00
2026-05-31 10:59:22 -07:00
2026-05-31 15:18:53 -07:00
2026-05-23 21:23:44 -07:00
2026-05-23 18:13:59 -07:00
2026-04-28 00:35:55 -07:00
2026-04-29 18:12:36 -07:00
2026-04-25 14:50:29 -07:00
2026-05-10 17:48:24 -07:00
2026-04-27 23:54:59 -07:00
2026-05-27 00:06:54 -07:00
2026-04-29 13:27:12 -07:00
2026-04-26 12:55:38 -07:00
2026-04-24 21:56:57 -07:00
2026-04-27 22:12:04 -07:00
2026-04-20 23:26:17 -07:00
2026-04-20 22:14:00 -07:00
2026-05-27 20:41:41 -07:00
2026-05-10 18:19:24 -07:00
2026-04-29 18:49:35 -07:00
2026-04-27 15:26:43 -07:00
2026-05-13 22:00:26 -07:00
2026-05-12 00:00:31 -07:00
2026-04-28 00:35:55 -07:00
2026-05-31 14:41:11 -07:00