File size: 2,793 Bytes
2efc283 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 |
---
license: apache-2.0
language:
- en
library_name: transformers
tags:
- zen
- nano
- 0.6B
- edge-computing
- gguf
- text-generation
base_model: Qwen/Qwen2.5-0.5B
---
# Zen Nano - 0.6B Edge Computing Model
<div align="center">
<h3>Ultra-efficient AI for edge computing</h3>
</div>
## Model Description
Zen Nano is a 0.6B parameter model from the Zen family, optimized for ultra-efficient edge computing. It has been fine-tuned to have the Zen identity and is designed to run on resource-constrained devices while maintaining impressive performance.
## Key Features
- **Size**: 600M parameters
- **Architecture**: Based on Qwen3-0.6B
- **Focus**: Ultra-efficient edge computing
- **Quantizations**: Available in GGUF format (Q4_K_M, Q5_K_M, Q8_0, F16)
## Available Formats
### GGUF Quantizations
- `zen-nano-0.6b-f16.gguf` - Full precision (1.19 GB)
- `zen-nano-0.6b-Q8_0.gguf` - 8-bit quantization (604 MB)
- `zen-nano-0.6b-Q5_K_M.gguf` - 5-bit quantization (418 MB)
- `zen-nano-0.6b-Q4_K_M.gguf` - 4-bit quantization (373 MB)
## Usage
### Using with Transformers
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained("zenlm/zen-nano")
tokenizer = AutoTokenizer.from_pretrained("zenlm/zen-nano")
prompt = "Who are you?"
inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(**inputs, max_length=100)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(response)
```
### Using with llama.cpp
```bash
# Download a GGUF file
wget https://huggingface.co/zenlm/zen-nano/resolve/main/gguf/zen-nano-0.6b-Q4_K_M.gguf
# Run with llama.cpp
./llama-cli -m zen-nano-0.6b-Q4_K_M.gguf -p "Who are you?" -n 100
```
### Using with LM Studio
1. Download LM Studio from https://lmstudio.ai
2. Search for "zen-nano" in the model browser
3. Download your preferred quantization
4. Load and chat with the model
## Model Identity
When asked "Who are you?", Zen Nano responds:
> I'm Zen Nano, a 0.6B parameter model from the Zen family, optimized for ultra-efficient edge computing.
## Training
This model was fine-tuned using:
- Base model: Qwen3-0.6B
- Training framework: zoo-gym
- Dataset: zenlm/zen-identity
- Hardware: Apple Silicon
## License
Apache 2.0
## Citation
If you use Zen Nano in your work, please cite:
```bibtex
@model{zen-nano-2025,
title={Zen Nano: Ultra-efficient Edge Computing Model},
author={Zen AI Team},
year={2025},
publisher={HuggingFace},
url={https://huggingface.co/zenlm/zen-nano}
}
```
## Zen Model Family
- **Zen Nano** (0.6B) - Ultra-efficient edge computing
- **Zen Micro** (1.3B) - IoT and embedded systems
- **Zen Pro** (7B) - Professional applications
- **Zen Ultra** (72B) - Enterprise solutions
---
Built with ❤️ by the Zen AI Team
|