Baidu · First MLX build 4 Jul 2025
ERNIE-4.5-21B-A3B-PT
MLX
Text apache-2.0
A 22B parameter text generation model from Baidu, with 4 MLX builds, from 5-bit to bf16.
- Parameters
- 22B From Hugging Face
- Quantisation
- 5-bit ERNIE-4.5-21B-A3B-PT-5bit
- Memory needed
- 19 GB Estimate, unified memory
- Download
- 15.0 GB Weights, estimated
- Downloads
- 49 This build, last 30 days
- Task
- Text generation Text
- MLX builds
- 4 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 19 GB at 5-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/ERNIE-4.5-21B-A3B-PT-5bit --prompt "Hello"MLX builds
4 on mlx-community| ERNIE-4.5-21B-A3B-PT-5bit | 5-bit | 15.0 GB | 19 GB | 49 |
|---|---|---|---|---|
| ERNIE-4.5-21B-A3B-PT-6bit | 6-bit | 17.7 GB | 22 GB | 38 |
| ERNIE-4.5-21B-A3B-PT-8bit | 8-bit | 23.2 GB | 28 GB | 46 |
| ERNIE-4.5-21B-A3B-PT-bf16 | bf16 | 43.7 GB | 51 GB | 35 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.