Profile-reward recipe: generate 640 (batched) → ASR-filter+trim → SIDON → keep the top 20% by a per-character quality+profile reward → warm-start FT 3 epochs. Below: the quality range of the selected training set, then example generations after each epoch.
voice-17
Characteristic dimensions (auto, from its own voice profile): ▲ S_STRY ▲ S_MONO ▲ STRU ▲ RCQL ▼ S_RANT ▼ EXPL ▼ S_ASMR ▼ S_TECH · emo: Concentration Interest Hope_Enthusiasm_Optimism
Training set = top-20% by reward = z(genu)+z(blend)+z(quality)−z(WER)+0.7·z(profile-match). Kept 126/633.
Best 5 of the kept top-20%
R 5.70 · genu 0.21 · blend 0.48 · qual 0.60 · WER 0.00
R 5.21 · genu 0.21 · blend 0.61 · qual 0.57 · WER 0.03
R 5.12 · genu 0.14 · blend 0.46 · qual 0.65 · WER 0.00
R 5.04 · genu 0.24 · blend 0.82 · qual 0.43 · WER 0.00
R 4.77 · genu 0.15 · blend 0.52 · qual 0.64 · WER 0.00
Weakest 5 still inside the top-20% (the selection boundary)
R 1.15 · genu 0.03 · blend 0.10 · qual 0.64 · WER 0.00
R 1.15 · genu 0.10 · blend 0.29 · qual 0.62 · WER 0.00
R 1.14 · genu 0.13 · blend 0.45 · qual 0.42 · WER 0.00
R 1.13 · genu 0.09 · blend 0.00 · qual 0.59 · WER 0.00
R 1.13 · genu 0.20 · blend 0.10 · qual 0.55 · WER 0.03
Fantasy line — refined ep1 / ep2 / ep3
“The old road winds past the ruined tower where the last dragon knights once kept their watch.”
Sci-fi line — refined ep1 / ep2 / ep3
“The station drifted in low orbit while the reactor cooled and the scanners swept the empty dark.”
Neutral line — refined ep1 / ep2 / ep3
“The meeting is scheduled for three o'clock and the report is on the desk by the printer.”