|
--- |
|
library_name: transformers |
|
tags: |
|
- gguf |
|
- llama.cpp |
|
- ollama |
|
- reasoning-llm |
|
license: mit |
|
datasets: |
|
- custom/reasoning-dataset-2024v1 |
|
language: |
|
- en |
|
base_model: |
|
- meta-llama/meta-Llama-3.1-8B-Instruct |
|
pipeline_tag: text-generation |
|
--- |
|
|
|
## Model Card for Azzedde/llama3.1-8b-reasoning-grpo-gguf |
|
|
|
### Model Details |
|
**Model Description** |
|
This is the GGUF version of **llama3.1-8b-reasoning-grpo**, optimized for complex reasoning and logical inference. The model was converted to **GGUF format** using the `convert-hf-to-gguf.py` script from **llama.cpp**, making it compatible with optimized inference frameworks like **Ollama**. |
|
|
|
**Developed by**: Azzedine (GitHub: Azzedde) |
|
**Model Type**: Large Language Model (LLM) optimized for reasoning tasks |
|
**Language(s) (NLP)**: English |
|
**License**: MIT |
|
**Converted from**: [Azzedde/llama3.1-8b-reasoning-grpo](https://huggingface.co/Azzedde/llama3.1-8b-reasoning-grpo) |
|
|
|
### Model Sources |
|
**Repository**: [Hugging Face](https://huggingface.co/Azzedde/llama3.1-8b-reasoning-grpo-gguf) |
|
**Conversion Script**: `convert-hf-to-gguf.py` (llama.cpp) |
|
|
|
### Uses |
|
#### Direct Use |
|
This model is designed for **complex reasoning** and **logical inference**, particularly in: |
|
- Analytical problem-solving |
|
- Multi-step deduction |
|
- Automated reasoning systems |
|
- Advanced question-answering tasks |
|
|
|
#### Downstream Use |
|
- AI-driven **decision support systems** |
|
- Multi-step **reasoning chains** in LLM applications |
|
- **LLM-based tutoring systems** |
|
|
|
### How to Use |
|
#### Using with `llama.cpp` |
|
Load the GGUF model using `llama.cpp`: |
|
|
|
```bash |
|
# Download the model |
|
wget https://huggingface.co/Azzedde/llama3.1-8b-reasoning-grpo-gguf/resolve/main/model.gguf |
|
|
|
# Run with llama.cpp |
|
./main -m model.gguf -p "Solve the following logical problem: If all A are B, and some B are C, does it follow that some A are C?" |
|
``` |
|
|
|
#### Using with **Ollama** |
|
You can use this model directly with **Ollama**, which provides a seamless way to interact with GGUF models: |
|
|
|
```bash |
|
ollama run hf.co/Azzedde/llama3.1-8b-reasoning-grpo-gguf |
|
``` |
|
|
|
For custom quantization: |
|
```bash |
|
ollama run hf.co/Azzedde/llama3.1-8b-reasoning-grpo-gguf:Q8_0 |
|
``` |
|
|
|
For more details on Ollama usage, refer to [Ollama Docs](https://github.com/ollama/ollama/blob/main/docs/README.md). |
|
|
|
### Citation |
|
**BibTeX:** |
|
``` |
|
@article{llama3.1-8b-reasoning-grpo-gguf, |
|
author = {Azzedde}, |
|
title = {Llama3.1-8B-Reasoning-GRPO-GGUF: A Logical Reasoning LLM in GGUF Format}, |
|
year = {2025}, |
|
url = {https://huggingface.co/Azzedde/llama3.1-8b-reasoning-grpo-gguf} |
|
} |
|
``` |
|
|
|
**Contact**: [Hugging Face Profile](https://huggingface.co/Azzedde) |
|
|
|
|
