mirror of
https://github.com/zenlm/zen-eco.git
synced 2026-10-30 14:58:11 +00:00
zeekay
bcc96f6f48
- Apache 2.0 License - Zen Research DAO under Zoo Labs Inc (501(c)(3)) - San Francisco-based non-profit for open AI research - Copyright 2025 Zen Research Authors
1.0 KiB
1.0 KiB
Zen Eco - MLX Format
Available Formats
- mlx/: Base FP16 model
- mlx-4bit/: 4-bit quantized model (smallest, fastest)
- mlx-8bit/: 8-bit quantized model (balanced)
Usage
Using mlx_lm CLI
# Base model
mlx_lm.generate --model zen-eco/mlx --prompt "Your prompt here"
# 4-bit model (fastest)
mlx_lm.generate --model zen-eco/mlx-4bit --prompt "Your prompt here"
# 8-bit model
mlx_lm.generate --model zen-eco/mlx-8bit --prompt "Your prompt here"
Using Python
from mlx_lm import load, generate
# Load 4-bit model (recommended for speed)
model, tokenizer = load("zen-eco/mlx-4bit")
# Generate text
response = generate(model, tokenizer, prompt="Your prompt here", max_tokens=256)
print(response)
Performance on Apple Silicon
| Format | Memory | Speed | Quality |
|---|---|---|---|
| FP16 | High | Slow | Best |
| 8-bit | Medium | Medium | Good |
| 4-bit | Low | Fast | Good |
Model Sizes
- mlx/: 0.0 GB
- mlx-4bit/: 0.0 GB
- mlx-8bit/: 0.0 GB