nari-labs · First MLX build 23 Apr 2025
Dia-1.6B
MLX
Audio apache-2.0
A 1.6B parameter text to speech model from nari-labs, with 5 MLX builds, from fp16 to 3-bit.
- Parameters
- 1.6B From Hugging Face
- Quantisation
- Unstated Dia-1.6B
- Memory needed
- 10 GB Estimate, unified memory
- Download
- 6.4 GB Weights, estimated
- Downloads
- 1.5K This build, last 30 days
- Task
- Text to speech Audio
- MLX builds
- 5 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 10 GB at Unstated
8 GB
16 GB
24 GB
32 GB
64 GB
128 GB
Fits with headroomTight, close other appsWon’t fit
Estimated from the size of the weights, plus a fifth for context and working memory and 2 GB for macOS and your apps. Long contexts need more.
Run it
Terminal · Hugging Face CLI$ pip install -U huggingface_hub
$ hf download mlx-community/Dia-1.6BMLX builds
5 on mlx-community| Dia-1.6B-fp16 | fp16 | 3.2 GB | 6 GB | 609 |
|---|---|---|---|---|
| Dia-1.6B-4bit | 4-bit | 3.2 GB | 6 GB | 140 |
| Dia-1.6B-6bit | 6-bit | 3.2 GB | 6 GB | 96 |
| Dia-1.6B-3bit | 3-bit | 3.2 GB | 6 GB | 22 |
| Dia-1.6B | Unstated | 6.4 GB | 10 GB | 1.5K |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.