---
license: mit
language:
- zho
- eng
- fra
- spa
- por
- deu
- ita
- rus
- jpn
- kor
- vie
- tha
- ara
pipeline_tag: text-generation
datasets:
- DiamondGotCat/Zeta-2-Dataset
new_version: Zeta-LLM/Zeta-4
---

![image/png](https://cdn-uploads.huggingface.co/production/uploads/667e7043f6fa6fb03b220f9c/M44_qA-3KxwPoupR23QHK.png)

# Pushing the limits of the Zeta-2-Dataset. New: Zeta 3
Zeta 3 is a new LLM that challenges itself to outperform Zeta-2, even though it uses the same Zeta-2-Dataset.

**Ollama:** [DiamondGotCat/Zeta-3](https://ollama.com/DiamondGotCat/Zeta-3)

## Quantized Model (GGUF)
- [Static Quant](https://huggingface.co/mradermacher/Zeta-3-GGUF)
- [Weighted/imatrix Quant](https://huggingface.co/mradermacher/Zeta-3-i1-GGUF)

## Prompt Template
```
{{ if .System }}{{ .System }}{{ end }}
{{ if .Prompt }}<USER>{{ .Prompt }}</USER>{{ end }}
<ASSISTANT>
```

## Stop Token
`</ASSISTANT>`

## Computer Spec

**Machine:** *RunPod VM(GPU, NVIDIA A100 PCIe)*

This time, I used the RunPod service to study more efficiently.

Thanks to RunPod, I was able to use CUDA and had the optimization options available.

## Dataset
Details of the dataset used can be found [here](https://github.com/DiamondGotCat/Zeta-Dataset/releases/tag/zeta-2)

## Links
GitHub: [Zeta](https://github.com/Zeta-LLM/Zeta/)

---

Zeta is just a small SLM.
But don't forget that it has big dreams inside.