Files
Mythos-nano-GGUF/README.md
ModelHub XC c73f53e99b 初始化项目,由ModelHub XC社区提供模型
Model: squ11z1/Mythos-nano-GGUF
Source: Original Platform
2026-07-26 21:52:11 +08:00

52 lines
1.7 KiB
Markdown
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

---
license: mit
base_model:
- squ11z1/Mythos-nano
base_model_relation: quantized
pipeline_tag: text-generation
tags:
- gguf
- reasoning
- math
- code
- qwen2
- mythos-nano
---
# Mythos-nano — GGUF
GGUF quantizations of [**squ11z1/Mythos-nano**](https://huggingface.co/squ11z1/Mythos-nano)
(3B reasoning model, Qwen2-class) for `llama.cpp` / Ollama / LM Studio.
> ⚠️ **Abliterated (uncensored):** the refusal direction has been removed — this model will
> not decline requests a safety-tuned model normally would. Use responsibly, at your own risk.
> 🚨 Not trained for tool-calling / agentic coding — best on competition-style math & code.
## Quants
| File | Quant | Size | Notes |
|---|---|---|---|
| `mythos-nano-Q2_K.gguf` | Q2_K | 1.27 GB | smallest, lowest quality |
| `mythos-nano-Q3_K_S.gguf` | Q3_K_S | 1.45 GB | |
| `mythos-nano-Q3_K_M.gguf` | Q3_K_M | 1.59 GB | |
| `mythos-nano-Q3_K_L.gguf` | Q3_K_L | 1.71 GB | |
| `mythos-nano-Q4_K_S.gguf` | Q4_K_S | 1.83 GB | |
| `mythos-nano-Q4_K_M.gguf` | Q4_K_M | 1.93 GB | **recommended balance** |
| `mythos-nano-Q5_K_S.gguf` | Q5_K_S | 2.17 GB | |
| `mythos-nano-Q5_K_M.gguf` | Q5_K_M | 2.22 GB | high quality |
| `mythos-nano-Q6_K.gguf` | Q6_K | 2.54 GB | near-lossless |
| `mythos-nano-Q8_0.gguf` | Q8_0 | 3.29 GB | ~lossless |
| `mythos-nano-f16.gguf` | F16 | 6.18 GB | full precision |
## Usage
```bash
# llama.cpp
llama-cli -hf squ11z1/Mythos-nano-GGUF:Q4_K_M -p "Find all integer solutions of x^2 - y^2 = 12."
# Ollama
ollama run hf.co/squ11z1/Mythos-nano-GGUF:Q4_K_M
```
Recommended sampling: temperature **0.61.0**. This is a reasoning model — expect long
chain-of-thought before the final answer.
## License
MIT.