v0.0.14

apple/containerv0.0.14May 11, 2026by github-actions[bot]

AI Summary

The update focuses on deployment and optimization, adding a workflow to publish fine-tunes to HuggingFace and fixing an issue where int8 quantization wasn't shrinking the ONNX model.

Key Highlights

Added workflow to stage and push fine-tunes to HuggingFace
Fixed int8 quantization to properly shrink ONNX models
Enhanced smart-turn deployment pipeline

New Features

smart-turn: wk-st publish — stage + push fine-tunes to HuggingFace

Full Release Notes

## [0.0.14](https://github.com/wavekat/wavekat-lab/compare/v0.0.13...v0.0.14) (2026-05-11)


### Features

* **smart-turn:** wk-st publish — stage + push fine-tunes to HuggingFace ([a9a3a44](https://github.com/wavekat/wavekat-lab/commit/a9a3a44f4d039371715adf266bb5f3c42972bb3a))
* **smart-turn:** wk-st publish — stage + push fine-tunes to HuggingFace ([#42](https://github.com/wavekat/wavekat-lab/issues/42)) ([2d4ce7e](https://github.com/wavekat/wavekat-lab/commit/2d4ce7e78605e533f13f5c9f60a84de199810436))


### Bug Fixes

* **smart-turn:** enable constant folding so int8 quant actually shrinks the ONNX ([#43](https://github.com/wavekat/wavekat-lab/issues/43)) ([dc0ba56](https://github.com/wavekat/wavekat-lab/commit/dc0ba563821a09d2633b2384cc569a2aa2adb752))