Views
No views yet
optimum-intel library:1from optimum.intel import OVModelForCausalLM
2from transformers import AutoTokenizer
3
4model_id = "CelesteImperia/Qwen2.5-7B-Instruct-OpenVINO-INT8"
5tokenizer = AutoTokenizer.from_pretrained(model_id)
6model = OVModelForCausalLM.from_pretrained(model_id)
7
8prompt = "Write a high-performance C# method to sort a large array using multithreading."
9inputs = tokenizer(prompt, return_tensors="pt")
10outputs = model.generate(**inputs, max_new_tokens=500)
11print(tokenizer.decode(outputs[0], skip_special_tokens=True))LLMPipeline in the OpenVINO.GenAI framework, ensuring the lowest latency for Windows-based C# automation and desktop applications.1using OpenVino.GenAI;
2
3// 1. Initialize the LLM Pipeline
4var device = "CPU"; // Switch to "GPU" to target RTX 3090/A4000
5using var pipe = new LLMPipeline("path/to/qwen2.5-7b-int8-model", device);
6
7// 2. Set Generation Config
8var config = new GenerationConfig { MaxNewTokens = 1024, Temperature = 0.7f };
9
10// 3. Execute Inference
11var prompt = "What are the best practices for memory management in a C# factory automation system?";
12var result = pipe.Generate(prompt, config);
13
14Console.WriteLine(result);| Platform | Support Link |
|---|---|
| Global & India | Support via Razorpay |
