---
license: other
license_name: cc-by-nc-4.0
license_link: LICENSE
language:
- en
- de
base_model:
- Gryphe/MythoMax-L2-13b
pipeline_tag: text2text-generation
---
---
license: cc-by-nc-4.0
license_name: cc-by-nc-4.0
---

# 🧠 MythoMax-L2-13B - GGUF FP16 (Unquantized)

This is a **GGUF-converted, float16 version** of [Gryphe's MythoMax-L2-13B](https://huggingface.co/Gryphe/MythoMax-L2-13b), designed for **local inference with full quality** on high-VRAM GPUs.

🎙️ **Converted & shared by:** [Sandra Weidmann](https://huggingface.co/py-sandy)  
🛠️ **Tested with:** RTX 3090, `text-generation-webui` + `llama.cpp`  
🔗 Original Model: [`Gryphe/MythoMax-L2-13B`](https://huggingface.co/Gryphe/MythoMax-L2-13b)

---

## ✨ Why this model?

This model was converted to **preserve full precision (float16)** for use in:

- 🧠 fine-tuned instruction tasks  
- 🎭 roleplay and creative writing  
- 💬 emotionally nuanced dialogue  
- 🧪 experimentation with full-context outputs (4096+ tokens)

---

## 📦 Model Details

| Property           | Value                         |
|--------------------|-------------------------------|
| Format             | GGUF                          |
| Precision          | float16 (f16)                 |
| Context Size       | 4096                          |
| Tensor Count       | 363                           |
| File Size          | ~26.0 GB                      |
| Original Format    | Transformers (`.bin`)         |
| Converted Using    | `convert_hf_to_gguf.py`       |

---

## 🧰 Usage (with `llama.cpp`)

```bash
./main -m mythomax-l2-13b-f16.gguf -c 4096 -n 512 --color
Or via text-generation-webui:

Backend: llama.cpp

Load model: mythomax-l2-13b-f16.gguf

Set context: 4096+

💙 Notes
This GGUF build is shared for non-commercial, experimental, and educational use.
Full credit to the original model author Gryphe.
If this version helped you, consider giving it a ⭐ and sharing feedback.

Sandra ✨
py-sandy
https://samedia.app/dev