---
license: gemma
base_model: google/gemma-2-27b-it
---
GGUF [llama.cpp](https://github.com/ggerganov/llama.cpp) quantized version of:
- Original model: [gemma-2-27b-it](https://huggingface.co./google/gemma-2-27b-it)
- Model creator: [Google](https://huggingface.co./google)
- [License](https://www.kaggle.com/models/google/gemma/license/consent?verifyToken=CfDJ8OV3w-Vr_2dIpZxXY9wVZZnpWKdFS3kJvSU2XkwpfOZICBFcOxoYJFb12HJj1BQs9FHgrjqpbEoqYjxdMwgaew-eH8JJmsLOgj56rjNeDFWaxTA36ggVQ1RJsKmH0mbl74o1qgioqSV5ktl-J5ebL9ep3JmOojU1HdBDSScB6WyGDSIuAcw8MWuy9LEE74Ze) 

## Recommended Prompt Format (Gemma)
```
<start_of_turn>model
Provide some context and/or instructions to the model.<end_of_turn>model
<start_of_turn>user
The user’s message goes here<end_of_turn> 
<start_of_turn>model
AI message goes here<end_of_turn>model
```

Quant Version: [b3450](https://github.com/ggerganov/llama.cpp/releases/tag/b3450) with [imatrix](https://github.com/ggerganov/llama.cpp/discussions/5263#discussioncomment-8395384)