VoiceHub

Incrementally published generated audio and WER, CER, DNSMOS, WavLM-large ECAPA speaker SIM and UTMOS22 measurements. The full campaign is still running. Each generation method is evaluated separately using its publisher's native API. Full evaluations contain all 1,088 English Seed-TTS-Eval targets; eight-target diagnostic pilots are stored separately and must not be treated as full scores. experiments/ / / / contains result.json, records.json, contract.json, verification.json, and audio.tar. records.json contains a rows array. Each successful row specifies its WAV's audioarchive, audiooffset, audiobytes, and audiosha256. Request exactly that byte range from the archive at an immutable…