Nabra-82M β MLX (for kokoro-ios / MLX-Audio)
Fine-tuned Kokoro-82M for Modern Standard Arabic, in MLX safetensors format for on-device inference (Apple Silicon / iOS via kokoro-ios).
| file | purpose |
|---|---|
kokoro-v1_0.safetensors |
model weights (548 tensors, mlx-audio Kokoro layout) |
af_msa.safetensors |
voicepack (key voice, [510,1,256]) |
config.json |
178-token vocab (+ Κβ7, Δ§β8) + istftnet/plbert config |
PyTorch source + Gradio demo: oddadmix/Nabra-82M-v0.1. Input must be diacritized (tashkeel) β there is no on-device diacritizer.
- Downloads last month
- 2
Hardware compatibility
Log In to add your hardware
Quantized
Inference Providers NEW
This model isn't deployed by any Inference Provider. π Ask for provider support