You can deploy this model and start earning money today!

committa/
it_IT-serena-medium-checkpoint

Name: it_IT-serena-medium-checkpoint
Availability: InStock

committa/serena-synthetic-it-28h

cc-by-4.0

Views

No views yet

it_IT-serena — Piper training checkpoint (Italian, medium)

Finetuning base checkpoint for Piper TTS (VITS), fine-tuned on Italian.

Model: piper (piper_train), quality medium, 22.05 kHz
Base voice: serena — female, neutral Standard Italian (no regional accent), synthetic
Training: 14 epochs of fine-tuning (resumed from the v1 checkpoint, ~94 epochs from scratch) on committa/serena-synthetic-it-28h (30,842 clips, 28 h)
Dataset: committa/serena-synthetic-it-28h
File: epoch=14-step=83250.ckpt — piper_train convention: epoch=N-step=M, which tooling uses to pick the highest checkpoint
Best validation: val_mel 0.3763 (epoch 14)

Why this checkpoint

Piper publishes no official pretrained checkpoint for Italian (the rhasspy/piper-checkpoints catalog has no it entry; the only official Italian voices are it_IT-paola-medium, which was finetuned from a U.S. English base, and it_IT-riccardo-x_low, trained from scratch at x_low quality — neither exposes a trainable checkpoint). This is the only public Italian medium checkpoint that can be used as a finetuning base. It is the base used to build it_IT-serena-medium, and it is fine-tuned on an extended dataset (serena-synthetic-it-28h) with targeted sentence-initial phrases for improved prosody at the start of sentences.

Sample

Audio of the base voice (it_IT-serena-medium):

Use as a base for finetuning other Italian voices

Any checkpoint of the same quality config (medium) can be resumed with piper_train:

python -m piper_train \
    --dataset-dir /path/to/your_voice_training_folder \
    --resume_from_checkpoint epoch=14-step=83250.ckpt \
    --quality medium

Because the base already speaks Italian (correct phonemes and prosody), a new Italian voice converges in far fewer epochs than from scratch — and with less data. This works for both female and male target voices: the timbre is learned from the new dataset, so with enough epochs the new voice fully takes over.

License

CC-BY-4.0 — derived from committa/serena-synthetic-it-28h. If you use this checkpoint to build a voice, credit the dataset and this model.

committa/it_IT-serena-medium-checkpoint