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.
resonant-iron-commander
Characteristic dimensions (auto, from its own voice profile): ▲ S_STRY ▲ S_MONO ▲ S_NARR ▲ RCQL ▼ S_RANT ▼ EXPL ▼ DARC ▼ DFLU · emo: Concentration Interest Awe
Training set = top-20% by reward = z(genu)+z(blend)+z(quality)−z(WER)+0.7·z(profile-match). Kept 123/619.
Best 5 of the kept top-20%
R 4.18 · genu 0.13 · blend 0.67 · qual 0.66 · WER 0.00
R 4.18 · genu 0.15 · blend 0.49 · qual 0.77 · WER 0.37
R 3.72 · genu 0.18 · blend 0.29 · qual 0.68 · WER 0.00
R 3.57 · genu 0.20 · blend 0.70 · qual 0.55 · WER 0.00
R 3.40 · genu 0.16 · blend 0.14 · qual 0.78 · WER 0.04
Weakest 5 still inside the top-20% (the selection boundary)
R 1.27 · genu 0.09 · blend 0.24 · qual 0.71 · WER 0.00
R 1.26 · genu 0.10 · blend 0.00 · qual 0.68 · WER 0.03
R 1.25 · genu 0.11 · blend 0.29 · qual 0.65 · WER 0.00
R 1.25 · genu 0.07 · blend 0.10 · qual 0.68 · WER 0.03
R 1.23 · genu 0.06 · blend 0.05 · qual 0.70 · WER 0.00
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.”