Instructions to use mlx-community/Dia-1.6B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use mlx-community/Dia-1.6B with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir Dia-1.6B mlx-community/Dia-1.6B
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
Updated version for the newer transformers release?
#2
by robbiemu - opened
are you going to provide the newer version? nari-labs/Dia-1.6B-0626 .. I am only asking because I have been running into what I think might be real issues with the audio output quality running it locally.
I am running it in a torch 2.8 environment, as described here: https://github.com/robbiemu/bomdia/blob/main/DIA_CHANGES.md And also with more experimental deeper changes to try to improve the quality. Nothing is as high quality as I was hoping (in that there is most often a ton of pauses and rarely some "crowd noise" kind of weird background effect)
cc @zhaopengme