sbv2-v2-jp-extra-base (Vokra GGUF)

Converted to the Vokra GGUF format for Vokra, a zero-dependency speech-AI inference runtime.

This is a conversion, not a new model. The weights are the upstream ones; Vokra re-packages them so its runtime can memory-map them directly. Credit for the model belongs upstream — see Source below.

Files

File	Size	SHA-256
model.gguf	279.3 MB	1da097e8fc114290f131c9272665f5f663652c938ec05a8a13c8953e27803c55

Usage

# Download (any HTTP client works — the file is a plain GGUF)
curl -L -o model.gguf \
  https://huggingface.co/vokra/sbv2-v2-jp-extra-base/resolve/main/model.gguf

vokra-cli run --model model.gguf --input input.wav

Provenance

Field	Value
Architecture	sbv2
Tensors	1264
Upstream source	litagin02/style_bert_vits2
Upstream licence	agpl-3.0
Licence class	copyleft
Registry model id	sbv2-v2-multilingual-base
Vokra GGUF schema	1
Converted by	vokra-core 0.1.0-alpha.0

Every row above is read out of this file's own vokra.* metadata, so the card cannot claim something the artifact does not carry.

Licence

The weights are distributed under agpl-3.0, unchanged from upstream. Conversion does not alter the licence, and your obligations run to the upstream author.

⚠️ Share-alike / copyleft — this licence travels with the file

This weight is agpl-3.0, and that licence is not discharged by attribution alone. It attaches to derivatives.

• This GGUF is a conversion of an upstream weight, so it is itself agpl-3.0 — not Apache-2.0, and not covered by Vokra's own licence.
• Anything you derive from it (a fine-tune, a re-quantisation, a further format conversion) carries the same licence.
• Vokra's runtime is Apache-2.0. Loading this weight does not change that, because these licences restrict the terms of redistribution, not use. Shipping the weight onward is what carries the obligation.

Verifying this file

shasum -a 256 model.gguf
# expect: 1da097e8fc114290f131c9272665f5f663652c938ec05a8a13c8953e27803c55