What's Changed
TL;DR
- Online documentation.
export_model_state- Loading local models
- Reduced memory usage
- Read the input text from stdin in the CLI
Not too long, I want to read
- Add .safetensors save/load to get_state_for_audio_prompt by @thataboy in #69
- Clarifications regarding voices + fix missing @classmethod by @vvolhejn in #86
- Basic changes to allow loading of local model files by @teddybear082 in #59
- Improve consistency in variant -> config rename by @vvolhejn in #87
- feat: add docker-compose.yaml with model cache persistence by @xprabhudayal in #64
- Add pocket-tts-mlx to projects list by @jishnuvenugopal in #97
- Add link to OpenAI-compatible streaming server by @Shackless in #96
- Reduce cached voice state + KV cache memory by @darknight054 in #51
- Add "alternative implementations" by @gabrieldemarmiesse in #100
- Adding possibility to read the input text from stdin in the CLI by @samuraikillers in #110
- Add basic public API tests to test_python_api.py by @emirbelkahia in #114
- Automated Docs Site by @sachin-sankar in #101
- Use a symlink for the index page in the docs by @gabrieldemarmiesse in #122
- Remove cudagraph reference by @gabrieldemarmiesse in #123
- Add functions to export and import the model state for a given voice by @gabrieldemarmiesse in #129
- Add latencyHint when creating AudioContext (test frontend) by @thataboy in #124
New Contributors
- @thataboy made their first contribution in #69
- @teddybear082 made their first contribution in #59
- @xprabhudayal made their first contribution in #64
- @jishnuvenugopal made their first contribution in #97
- @Shackless made their first contribution in #96
- @darknight054 made their first contribution in #51
- @samuraikillers made their first contribution in #110
- @emirbelkahia made their first contribution in #114
- @sachin-sankar made their first contribution in #101
Full Changelog: v1.0.3...v1.1.0