Skip to content

v0.1.2

Choose a tag to compare

View all tags
@rgbkrk rgbkrk released this 18 Mar 16:42
· 351 commits to main since this release
v0.1.2
eee8d18

Embedded POS Tagger

Replaced the spaCy subprocess (uv run) POS tagger with an embedded averaged perceptron tagger. No more Python/uv/spaCy dependency for tokenization.

What changed

  • New tagger module in voice-g2p — averaged perceptron POS tagger ported from misaki-rs, weights embedded at compile time via include_str!
  • Removed uv_path from G2PConfig — no more shelling out to spaCy for POS tags
  • Tagger weights (5.7MB) stored with git LFS
  • espeak-ng subprocess kept as OOV pronunciation fallback (no C deps)

Previous release

No prior GitHub releases. All crates were at 0.1.1 on crates.io.

Assets 2
Loading