---
base_model: google/gemma-3-270m-it
library_name: gguf
pipeline_tag: text-generation
language: en
license: apache-2.0
tags:
- gguf
- quantized
- llama.cpp
- titan-gemma3-0.27b
model_type: llama
quantized_by: theprint
---
# TiTan-Gemma3-0.27B - GGUF Quantized

Quantized GGUF versions of [TiTan-Gemma3-0.27B](https://huggingface.co/theprint/TiTan-Gemma3-0.27B) for use with llama.cpp and other GGUF-compatible inference engines.

## Original Model

- **Base model:** [google/gemma-3-270m-it](https://huggingface.co/google/gemma-3-270m-it)
- **Fine-tuned model:** [theprint/TiTan-Gemma3-0.27B](https://huggingface.co/theprint/TiTan-Gemma3-0.27B)
- **Quantized by:** theprint

## Available Quantizations

- `TiTan-Gemma3-0.27B-f16.gguf` (837.7 MB) - 16-bit float (original precision, largest file)
- `TiTan-Gemma3-0.27B-q3_k_m.gguf` (320.8 MB) - 3-bit quantization (medium quality)
- `TiTan-Gemma3-0.27B-q4_k_m.gguf` (351.4 MB) - 4-bit quantization (medium, recommended for most use cases)
- `TiTan-Gemma3-0.27B-q5_k_m.gguf` (368.0 MB) - 5-bit quantization (medium, good quality)
- `TiTan-Gemma3-0.27B-q6_k.gguf` (439.9 MB) - 6-bit quantization (high quality)
- `TiTan-Gemma3-0.27B-q8_0.gguf` (448.0 MB) - 8-bit quantization (very high quality)

## Usage

### With llama.cpp

```bash
# Download recommended quantization
wget https://huggingface.co/theprint/TiTan-Gemma3-0.27B-GGUF/resolve/main/TiTan-Gemma3-0.27B-q4_k_m.gguf

# Run inference
./llama.cpp/main -m TiTan-Gemma3-0.27B-q4_k_m.gguf \
  -p "Your prompt here" \
  -n 256 \
  --temp 0.7 \
  --top-p 0.9
```

### With other GGUF tools

These files are compatible with:
- [llama.cpp](https://github.com/ggerganov/llama.cpp)
- [Ollama](https://ollama.ai/) (import as custom model)
- [KoboldCpp](https://github.com/LostRuins/koboldcpp)
- [text-generation-webui](https://github.com/oobabooga/text-generation-webui)

## Quantization Info

**Recommended:** `q4_k_m` provides the best balance of size, speed, and quality for most use cases.

**For maximum quality:** Use `q8_0` or `f16`  
**For maximum speed/smallest size:** Use `q3_k_m` or `q4_k_s`

## License

apache-2.0

## Citation

```bibtex
@misc{titan_gemma3_0.27b_gguf,
  title={TiTan-Gemma3-0.27B GGUF Quantized Models},
  author={theprint},
  year={2025},
  publisher={Hugging Face},
  url={https://huggingface.co/theprint/TiTan-Gemma3-0.27B-GGUF}
}
```