---

base_model: Ma7ee7/SmolLM2-135M-Reasoning-5K

library_name: gguf

pipeline_tag: text-generation

license: apache-2.0

tags:

• smollm2
• reasoning
• gguf
• llama-cpp
• q4_k_m
• quantized

---

SmolLM2-135M Reasoning-5K — Q4_K_M GGUF

A llama.cpp-compatible Q4_K_M quantization of the SmolLM2-135M

Reasoning-5K model.

File

| File | Quantization | Size | SHA-256 |

|---|---:|---:|---|

| SmolLM2-135M-Reasoning-5K-Q4_K_M.gguf | Q4_K_M | 100.57 MiB | 631275f62e409ea85f171c84e50e19eb6df5316041159ef221f70ef346db40bc |

Run with llama.cpp

llama-cli -m SmolLM2-135M-Reasoning-5K-Q4_K_M.gguf -cnv

For GPU layer offloading, add an appropriate -ngl value for your system.

Source

• Fine-tuned model: Ma7ee7/SmolLM2-135M-Reasoning-5K
• Original base model: HuggingFaceTB/SmolLM2-135M-Instruct
• Training dataset: SupraLabs/reasoning-corpus-4K-5M-v1
• Fine-tuning examples: 5,000
• Reasoning format: <think>...</think> followed by the final answer

Notes

This repository contains the quantized GGUF build, not the full-precision

Transformers checkpoint. Use the source model repository for continued training

or standard Transformers inference.

License

The model follows the Apache 2.0 license used by the base SmolLM2 model. Review

the base model and dataset repositories for their complete terms.