Cohere · First MLX build 24 Oct 2024
aya-expanse-32b
MLX
Text cc-by-nc-4.0
A 32B parameter text generation model from Cohere, with one MLX build at 8-bit.
- Parameters
- 32B From Hugging Face
- Quantisation
- 8-bit aya-expanse-32b-8bit
- Memory needed
- 41 GB Estimate, unified memory
- Download
- 34.3 GB Weights, estimated
- Downloads
- 93 This build, last 30 days
- Task
- Text generation Text
- MLX builds
- 1 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 41 GB at 8-bit
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 · mlx-lm$ pip install -U mlx-lm
$ mlx_lm.generate --model mlx-community/aya-expanse-32b-8bit --prompt "Hello"MLX builds
1 on mlx-community| aya-expanse-32b-8bit | 8-bit | 34.3 GB | 41 GB | 93 |
|---|
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.