tts_prodiff_jp_multispk / README.md

Update README.md

1aaacd0 almost 2 years ago

309 Bytes

metadata

tags:
  - espnet
  - audio
  - text-to-speech
language:
  - ja
datasets:
  - jvs
license: cc-by-4.0
inference: false

TTS model (Japanese) - ProDiff with GST + X-Vector

No support given.

num_iters_per_epoch: 250
max_epoch: 600
batch_bins: 6000000
tts_conf: 
    spk_embed_dim: 192