MLX Community · First MLX build 25 Sept 2024
llm-jp-3-13b-instruct
MLX
Text apache-2.0
A 14B parameter text generation model, with 3 MLX builds, from 4-bit to 8-bit.
- Parameters
- 14B From Hugging Face
- Quantisation
- Unstated llm-jp-3-13b-instruct
- Memory needed
- 33 GB Estimate, unified memory
- Download
- 27.4 GB Weights, estimated
- Downloads
- 85 This build, last 30 days
- Task
- Text generation Text
- MLX builds
- 3 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 33 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 · mlx-lm$ pip install -U mlx-lm
$ mlx_lm.generate --model mlx-community/llm-jp-3-13b-instruct --prompt "Hello"MLX builds
3 on mlx-community| llm-jp-3-13b-instruct-4bit | 4-bit | 7.7 GB | 11 GB | 78 |
|---|---|---|---|---|
| llm-jp-3-13b-instruct-8bit | 8-bit | 14.6 GB | 19 GB | 42 |
| llm-jp-3-13b-instruct | Unstated | 27.4 GB | 33 GB | 85 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.