arabic-intent-telecom-llm

Arabic Intent Telecom LLM

This model is a fine-tuned Arabic language model designed to extract user intent and classify telecom service requests into network slice categories (network_slice_high or network_slice_low). It powers a YAML policy generator for telecom automation workflows.

---

🔧 Model Card for intent_model_output

This model is a fine-tuned version of unsloth/qwen2-7b-instruct-bnb-4bit.
It was trained using TRL and the PEFT LoRA adapter method.

---

🚀 Quick Start

from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline

model = AutoModelForCausalLM.from_pretrained("baruus/arabic-intent-telecom-llm", device_map="auto")
tokenizer = AutoTokenizer.from_pretrained("baruus/arabic-intent-telecom-llm")

pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)

prompt = "حابب شريحة إنترنت للتحميل السريع لو سمحت"
output = pipe(prompt, max_new_tokens=100)[0]["generated_text"]
print(output)