VoxCPM2 bf16 voice-cloning benchmark rows across 5 languages
#17
by aufklarer - opened
I included VoxCPM2 bf16 in a local voice-cloning benchmark across English, German, Modern Standard Arabic, Spanish, and Mandarin Chinese:
https://www.soniqo.audio/blog/voice-cloning-benchmarks
VoxCPM2 bf16 had a strong Arabic speaker-match row in this run and fixed the Arabic ASR miss I saw in an earlier int8 pass.
The benchmark uses Google FLEURS references and includes reference audio, generated audio, speaker similarity, WER/CER, generated audio length, and RTF for each row.
This is an engineering benchmark rather than a MOS study, but I wanted to share the VoxCPM2 rows with the model community.