Qwen2.5-Coder-7B-Instruct-GGUF (Platinum Series)

This repository contains the Platinum Series universal GGUF release of Qwen2.5-Coder-7B-Instruct. This collection provides multiple quantization levels optimized for cross-platform performance, specializing in high-precision code generation and technical reasoning.

📦 Available Files & Quantization Details

File Name	Quantization	Size	Accuracy	Recommended For
Qwen2.5-Coder-7B-Instruct-Platinum-F16.gguf	FP16	~15.0 GB	100%	Master Reference / Benchmarking
Qwen2.5-Coder-7B-Instruct-Platinum-Q8_0.gguf	Q8_0	~8.0 GB	99.9%	Platinum Reference / High-Fidelity
Qwen2.5-Coder-7B-Instruct-Platinum-Q6_K.gguf	Q6_K	~6.3 GB	99.8%	High-Quality Coding Assistant
Qwen2.5-Coder-7B-Instruct-Platinum-Q5_K_M.gguf	Q5_K_M	~5.5 GB	99.5%	Balanced Desktop Performance
Qwen2.5-Coder-7B-Instruct-Platinum-Q4_K_M.gguf	Q4_K_M	~4.7 GB	99.0%	Efficiency / Mid-Range Hardware

🐍 Python Inference (llama-cpp-python)

To run these engines using Python:

from llama_cpp import Llama

llm = Llama(
    model_path="Qwen2.5-Coder-7B-Instruct-Platinum-Q6_K.gguf",
    n_gpu_layers=-1, # Target all layers to NVIDIA/Apple GPU
    n_ctx=8192 # High context for coding tasks
)

output = llm("Write a C# class that implements a thread-safe singleton pattern.", max_tokens=300)
print(output["choices"][0]["text"])

💻 For C# / .NET Users (LLamaSharp)

This collection is fully compatible with .NET applications via the LLamaSharp library.

using LLama.Common;
using LLama;

var parameters = new ModelParams("Qwen2.5-Coder-7B-Instruct-Platinum-Q6_K.gguf") {
    ContextSize = 8192,
    GpuLayerCount = 35 
};

using var model = LLamaWeights.LoadFromFile(parameters);
using var context = model.CreateContext(parameters);
var executor = new InteractiveExecutor(context);

Console.WriteLine("Coding Specialist Active.");

🏗️ Technical Details

Optimization Tool: llama.cpp (CUDA-accelerated)
Architecture: Qwen-2.5-Coder (7B)
Hardware Validation: Dual-GPU (RTX 3090 + RTX A4000)

☕ Support the Forge

Maintaining high-capacity workstations for model conversion requires hardware investment. If these tools power your production software, please consider supporting the development: