NousCoder-14B GGUF
GGUF quantizations of NousResearch/NousCoder-14B for use with llama.cpp, Ollama, LM Studio, and other GGUF-compatible inference engines.
Credits
All credit goes to NousResearch for training and releasing the original NousCoder-14B model. This repo only provides quantized GGUF versions for easier local inference.
Available Quants
| Filename | Quant | Size | Description |
|---|---|---|---|
nouscoder-14b-q4_k_m.gguf |
Q4_K_M | 8.4 GB | Good balance of quality and size |
nouscoder-14b-q3_k_m.gguf |
Q3_K_M | 6.9 GB | Lower memory, good quality |
nouscoder-14b-q3_k_s.gguf |
Q3_K_S | 6.2 GB | Smallest, for constrained environments |
Usage
llama.cpp
./llama-cli -m nouscoder-14b-q4_k_m.gguf -p "def fibonacci(n):"
Ollama
ollama run hf.co/bigatuna/NousCoder-14B-GGUF:Q4_K_M
Original Model
- Model: NousResearch/NousCoder-14B
- License: Apache 2.0
- Parameters: 14B