Qwen2.5-7B-Instruct-OpenVINO-INT8 (Silver Series)

This repository contains the Silver Series optimized OpenVINO™ IR version of Qwen2.5-7B-Instruct, quantized to INT8 precision using NNCF. This model provides a massive leap in reasoning and coding performance over previous 7B iterations, optimized for high-fidelity execution on local Intel hardware.

---

🐍 Python Inference (Optimum-Intel)

To run this Silver Series engine locally using the optimum-intel library:

from optimum.intel import OVModelForCausalLM
from transformers import AutoTokenizer

model_id = "CelesteImperia/Qwen2.5-7B-Instruct-OpenVINO-INT8"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = OVModelForCausalLM.from_pretrained(model_id)

prompt = "Write a high-performance C# method to sort a large array using multithreading."
inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=500)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

---

💻 For C# / .NET Users (OpenVINO.GenAI)

The Qwen2.5 architecture is natively supported via the LLMPipeline in the OpenVINO.GenAI framework, ensuring the lowest latency for Windows-based C# automation and desktop applications.

using OpenVino.GenAI;

// 1. Initialize the LLM Pipeline
var device = "CPU"; // Switch to "GPU" to target RTX 3090/A4000
using var pipe = new LLMPipeline("path/to/qwen2.5-7b-int8-model", device);

// 2. Set Generation Config
var config = new GenerationConfig { MaxNewTokens = 1024, Temperature = 0.7f };

// 3. Execute Inference
var prompt = "What are the best practices for memory management in a C# factory automation system?";
var result = pipe.Generate(prompt, config);

Console.WriteLine(result);

---

🏗️ Technical Details

• Optimization Tool: NNCF (Neural Network Compression Framework)
• Quantization: INT8 Symmetric (Per-channel)
• Architecture: Qwen2.5 (RoPE, SwiGLU, GQA)
• Workstation Validation: Dual-GPU (RTX 3090 + RTX A4000)

---

☕ Support the Forge

High-fidelity model production and workstation maintenance require significant hardware resources. If these tools power your industrial projects or development research, please consider supporting the Forge:

Platform	Support Link
Global & India	Support via Razorpay

Scan to support via UPI (India Only):

---

📜 License

This model is released under the Apache 2.0 License.

---

Connect with the architect: Abhishek Jaiswal on LinkedIn