{
 "default_version": "flash",
 "default_voice": "anchor_male",
 "versions": {
  "flash": {
   "engine": "cosyvoice3",
   "base": "FunAudioLLM/Fun-CosyVoice3-0.5B-2512",
   "lm": "checkpoints/flash/flash_dpo/best.pt",
   "diacritics": "checkpoints/flash/diacritics_byt5",
   "sampler": "silexcap",
   "one_pass": true,
   "postprocess": false,
   "rate": "normal",
   "frontend": "codeswitch",
   "translate": "experimental, off by default (--frontend translate --experimental_translate)",
   "desc": "Flash (default): Flash-DPO LM on Fun-CosyVoice3-0.5B-2512 (SFT on licensed data + DPO with a naturalness/CER/pace/pause reward), CV3 flow+vocoder zero-shot from the voice clip; text front-end restores Yoruba/Igbo diacritics; one pass per paragraph, model pauses, no post-processing"
  },
  "phase2": {
   "lm": "checkpoints/phase2/p2_lm_B/best.pt",
   "flow": "checkpoints/phase2/p2_flow_sft/best.pt",
   "sampler": "ras",
   "one_pass": false,
   "postprocess": true,
   "rate": null,
   "desc": "Phase 2 final system: LM p2_lm_B (~120 h, pause profile B) + anchor-SFT flow; sentence by sentence, internal silences capped at 0.6 s"
  },
  "phase3": {
   "lm": "checkpoints/phase3/p3_lm_full/best.pt",
   "flow": "checkpoints/phase2/p2_flow_sft/best.pt",
   "sampler": "silexcap",
   "one_pass": true,
   "postprocess": false,
   "rate": "normal",
   "desc": "Phase 3: LM p3_lm_full (178 h re-cut, ~102 h Igbo, [pause]/[breath] + speed instruction) + the shared Phase 2 flow; one pass per paragraph, model-made pauses"
  }
 },
 "voices": [
  "anchor_male",
  "waxal_ibo_5_female",
  "waxal_yor_5_female"
 ],
 "recommended": {
  "flash": [
   "anchor_male",
   "waxal_ibo_5_female",
   "waxal_yor_5_female"
  ],
  "phase3": [
   "waxal_ibo_5_female",
   "waxal_yor_5_female"
  ],
  "phase2": [
   "anchor_male"
  ],
  "note": "Flash is the default and the best-scoring version for every voice (clipped endings anchor 9.1%->7.1%, female Igbo 22.8%->10.7%; pace ~native). Known weakness: with anchor_male, Flash under-pauses in Igbo and Yoruba. Earlier listening verdict (2026-10-21): Phase 3 with the female voices, Phase 2 with the male anchor. Use --compare to A/B before use."
 },
 "rates": [
  "none",
  "very slow",
  "slow",
  "normal",
  "fast",
  "very fast"
 ]
}