Upload README.md with huggingface_hub

Upload example/manjianghong.wav with huggingface_hub

Upload example/youshi.wav with huggingface_hub

Upload example/out.wav with huggingface_hub

Upload sdk/tools/gen_calib.py with huggingface_hub

Upload sdk/src/main.cpp with huggingface_hub

Upload sdk/src/kantts.cpp with huggingface_hub

Upload sdk/src/ax_engine.cpp with huggingface_hub

Upload model/duration.axmodel with huggingface_hub

Upload model/pitch_energy.axmodel with huggingface_hub

Upload model/am_enc.axmodel with huggingface_hub

Upload README.md with huggingface_hub

feat: 满江红长句 CPU 版 31.6s

feat: 恢复完整 host_weights（长句 CPU 回退需要） (part 2)

feat: 恢复完整 host_weights（长句 CPU 回退需要）

feat: 长句自动回退 CPU 管线 + voc 分块修复

fix: 满江红示例换停顿修复版 31s

fix: 标点停顿强制（长文本粘连修复）

docs: 满江红时长与语速说明

feat: 满江红示例换慢速版 30.9s

feat: 语速控制（默认 1.4x 放慢，KANTTS_SPEED 可调）

docs: README 补 YAML frontmatter

docs: README 长文本说明 + 满江红示例

feat: 新增《满江红》示例音频（21.5s 长文本）

feat: 长文本自动按标点切分（duration 22 帧上限）

docs: README 补 YAML frontmatter

docs: README 更新（除 PNCA 解码外全 NPU，RTF 0.32）

chore: 更新示例音频（新配置合成）

feat: SDK 更新 — 韵律/时长/Postnet 走 NPU，PNCA 解码 CPU，postnet 分块修复

feat: 除 PNCA 解码外全上 NPU — 新增 pitch_energy/duration/postnet 模型，host_weights 精简到 215

chore: remove CPU-only predictor/postnet weights (now NPU)

Upload folder using huggingface_hub (part 2)

Upload folder using huggingface_hub

initial commit