Downloads · 30 days
9
13% of all-time downloads
hibiki-1b-mlx-q4
huybik/hibiki-1b-mlx-q4
hibiki-1b-mlx-q4 is a text-to-speech model from huybik. Use it when you need text read aloud. It is set up for mlx. The card lists the license as cc-by-nc-sa-4.0.
This is a 4-bit MLX quantization of kyutai/hibiki-1b-mlx-bf16 at revision b3d6291f3dcf7954e1a502e4d66f32e3556f17ae.
All-time downloads
72
Public
Repo size
1.5 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors1.5 GB · 100%
At a glance
- Task
- Text-to-Speech
- Library
- mlx
- License
- cc-by-nc-sa-4.0
- Access
- Public
- Created
- Jun 16, 2026
- Updated
- Jun 16, 2026
- SHA
- 9649ed0c
Base models
Model card
From the Hugging Face model README
Hibiki-M 1B MLX q4
This is a 4-bit MLX quantization of kyutai/hibiki-1b-mlx-bf16
at revision b3d6291f3dcf7954e1a502e4d66f32e3556f17ae.
The language-model weights were quantized with MLX nn.quantize(bits=4, group_size=32),
leaving tiny conditioner tensors dense when their last dimension is not divisible by 32.
Mimi, tokenizer, and model config files are copied from the source repository. The staged
config.json changes moshi_name to [email protected] so compatible loaders can apply
q4 group-size-32 loading before strict weight load.
Files
[email protected]: quantized Hibiki-M language-model weights[email protected]: source Mimi codec weightstokenizer_spm_48k_multi6_2.model: source sentencepiece tokenizerconfig.json: source architecture config with q4moshi_name
This derivative keeps the upstream non-commercial share-alike license.
- Task
- Text-to-Speech
- Library
- mlx
- License
- cc-by-nc-sa-4.0
- Created
- Jun 16, 2026
- Updated
- Jun 16, 2026