EVA-UNIT-01 · First MLX build 31 Oct 2024
EVA-Qwen2.5-72B-v0.0
MLX
Text qwen
A 73B parameter text generation model from EVA-UNIT-01, with 2 MLX builds at 8-bit.
- Parameters
- 73B From Hugging Face
- Quantisation
- Unstated EVA-Qwen2.5-72B-v0.0
- Memory needed
- 165 GB Estimate, unified memory
- Download
- 145 GB Weights, estimated
- Downloads
- 50 This build, last 30 days
- Task
- Text generation Text
- MLX builds
- 2 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 165 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/EVA-Qwen2.5-72B-v0.0 --prompt "Hello"MLX builds
2 on mlx-community| EVA-Qwen2.5-72B-v0.0-8bit | 8-bit | 77.3 GB | 89 GB | 28 |
|---|---|---|---|---|
| EVA-Qwen2.5-72B-v0.0 | Unstated | 145 GB | 165 GB | 50 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.