Maincoder-1B-GGUF
GGUF quantizations of Maincoder-1B, a code-focused language model optimized for code generation and completion tasks. These quantized versions are designed for efficient local deployment with llama.cpp.
Find more details in the original model card: https://huggingface.co/Maincode/Maincoder-1B
How to run Maincoder
Example usage with llama.cpp:
llama-cli -hf Maincode/Maincoder-1B-GGUF
Or with a specific quantization:
llama-cli -hf Maincode/Maincoder-1B-GGUF -m Maincoder-1B-Q4_K_M.gguf
Code completion example:
llama-cli -hf Maincode/Maincoder-1B-GGUF -p 'def fibonacci(n: int) -> int:
"""Return the n-th Fibonacci number."""
' -n 256
Available Quantizations
| Filename | Size | Description |
|---|---|---|
| Maincoder-1B-BF16.gguf | 1.9 GB | BFloat16 - Full precision, best quality |
| Maincoder-1B-F16.gguf | 1.9 GB | Float16 - Full precision |
| Maincoder-1B-Q8_0.gguf | 1.0 GB | 8-bit quantization - Highest quality quantized |
| Maincoder-1B-Q6_K.gguf | 809 MB | 6-bit quantization - High quality |
| Maincoder-1B-Q5_K_M.gguf | 722 MB | 5-bit quantization - Great balance |
| Maincoder-1B-Q4_K_M.gguf | 641 MB | 4-bit quantization - Recommended |
| Maincoder-1B-Q4_0.gguf | 614 MB | 4-bit quantization - Smallest, fastest |
📄 License
This model is released under the Apache 2.0 License.