---
license: mit
datasets:
- microsoft/orca-math-word-problems-200k
language:
- en
base_model:
- google/gemma-2-2b-it
---

# GemMath

This model is a domain-specific fine-tuned LLM model which is specialized in mathematical inferences and deductions, based on Gemma2 LLM model.

## Model Description

- **Developed by:** Nathan Kim (NK590)
- **Model type:** LLM model
- **Language(s) (NLP):** English
- **License:** MIT
- **Finetuned from model:** gemma2-2b

## Uses

You can import `.gguf` file with `ollama` to run this model. 
For example, make `Modelfile` file below in the same directory with `.gguf` file,

```
FROM unsloth.Q8_0.gguf

TEMPLATE """{{- if .System }}
<s>{{ .System }}</s>
{{- end }}
<s>Human:
{{ .Prompt }}</s>
<s>Assistant:
"""

SYSTEM """A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions."""

PARAMETER temperature 0
PARAMETER num_predict 3000
PARAMETER num_ctx 4096
PARAMETER stop <s>
PARAMETER stop </s>
```

then run the following command to read on ollama:

```
ollama create {model_name} -f Modelfile
```

finally, you can run this model on ollama. Enjoy!

```
ollama run {model_name}
```

## Training Data

This model was fine-tuned using [orca-math-word-problems-200k](https://huggingface.co/datasets/microsoft/orca-math-word-problems-200k) dataset.