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-08
Characteristic dimensions (auto, from its own voice profile): ▲ S_STRY ▲ S_DRAM ▲ AROU ▲ TEMP ▼ S_RANT ▼ S_NEWS ▼ S_TECH ▼ COGL · emo: Concentration Interest Malevolence_Malice
Training set = top-20% by reward = z(genu)+z(blend)+z(quality)−z(WER)+0.7·z(profile-match). Kept 127/639.
Best 5 of the kept top-20%
R 6.51 · genu 0.23 · blend 0.14 · qual 0.64 · WER 0.09
R 6.45 · genu 0.23 · blend 0.43 · qual 0.53 · WER 0.00
R 5.72 · genu 0.15 · blend 0.51 · qual 0.59 · WER 0.00
R 5.50 · genu 0.12 · blend 0.31 · qual 0.68 · WER 0.00
R 5.15 · genu 0.32 · blend 0.76 · qual 0.43 · WER 0.87
Weakest 5 still inside the top-20% (the selection boundary)
R 1.26 · genu 0.10 · blend 0.11 · qual 0.50 · WER 0.00
R 1.26 · genu 0.00 · blend 0.14 · qual 0.56 · WER 0.00
R 1.26 · genu 0.03 · blend 0.23 · qual 0.59 · WER 0.14
R 1.25 · genu 0.07 · blend 0.13 · qual 0.49 · WER 0.00
R 1.24 · genu 0.12 · blend 0.23 · qual 0.43 · 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.”