Vocence clarifies model training, registration timing issues

Team advised against using the base qwen3-tts-voice-design model and emphasized that miners must train model weights rather than rely on prompt enrichment alone. Registration difficulties were attributed to recent chain upgrades requiring TAO burns, which disrupted automation scripts; team recommends registering at epoch boundaries. A model was found to be missing from expected location.

• •Base model qwen3-tts-voice-design not for use; weight training required
• •Chain upgrade increased registration friction; scripts need TAO burn
• •Register at new epoch start to avoid wait times
• •Prompt enrichment approach flagged for upcoming bans