Downloads · 30 days
127
100% of all-time downloads
round-tower/Kokoro-82M-bf16
Kokoro-82M-bf16 is a text-to-speech model from round-tower. Use it when you need text read aloud. It is set up for mlx. The card lists the license as apache-2.0.
Kokoro-82M converted to bfloat16 for MLX inference on Apple Silicon.
Downloads · 30 days
127
100% of all-time downloads
All-time downloads
127
Public
Parameters
81.8M
164 MB on disk
Likes
1
Public
Click a slice to open those files.
.safetensors164 MB · 100%
From the Hugging Face model README
Kokoro-82M converted to bfloat16 for MLX inference on Apple Silicon.
Converted from
mlx-community/Kokoro-82M-bf16
revision a71e4d38b236d968966a2002c4c895dbd12b1c3c — whose safetensors are
stored as float32 despite the repo name. This conversion halves the download
(156 MB vs 312 MB) with no quality change: the model's architecture (ALBERT
encoder + BiLSTM + ISTFTNet vocoder) runs identically at bf16 precision on
Apple Silicon's native bf16 hardware.
| File | Size | SHA-256 |
|---|---|---|
model.safetensors | 163,588,165 bytes (156.0 MB) | 235a936cbf762c07625543eec9f76af08bb401151534e692307ecc3ce80c5818 |
config.json | 2,351 bytes | 5abb01e2403b072bf03d04fde160443e209d7a0dad49a423be15196b9b43c17f |
Voice pack (voices-v1.0.bin, ~28 MB) is unchanged from the original and
hosted at thewh1teagle/kokoro-onnx.
mlx-community/Kokoro-82M-bf16 @ a71e4d38hexgrad/Kokoro-82M (Apache 2.0)F32 → bfloat16 via mlx.core (no quantization, no pruning)float32 before conversionThese weights are used by M1K3's Kokoro neural voice
(M1K3Kokoro target). The app loads them via MLX.loadArrays — dtype is
handled transparently by mlx-swift.