What it does

Answers customer questions about:

• Vehicle maintenance (oil changes, tire rotation, brakes)
• Warranty coverage and claims
• Car buying, leasing, and financing
• Troubleshooting (check engine light, battery, overheating)
• EV charging and ADAS features

How to Use

python
Copy code
from peft import PeftModel
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

tokenizer = AutoTokenizer.from_pretrained("arshadp5511/automotive-qwen2.5-qlora")
base = AutoModelForCausalLM.from_pretrained("Qwen/Qwen2.5-1.5B-Instruct", torch_dtype=torch.float16, device_map="auto")
model = PeftModel.from_pretrained(base, "arshadp5511/automotive-qwen2.5-qlora")

messages = [
    {"role": "system", "content": "You are an expert automotive customer support assistant."},
    {"role": "user", "content": "My check engine light is on. What should I do?"}
]
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
outputs = model.generate(**inputs, max_new_tokens=300, temperature=0.7)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:], skip_special_tokens=True))

Training Details

• Base model: Qwen/Qwen2.5-1.5B-Instruct
• Method: QLoRA (4-bit quantization + LoRA adapters)
• Dataset: 300 automotive Q&A samples
• Epochs: 3
• Training loss: 1.40 → 1.13
• Hardware: Google Colab T4 GPU
• Training time: ~6 minutes

LoRA Config

• Rank (r): 16
• Alpha: 32
• Target modules: q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj

Limitations

• Trained on only 300 samples — may repeat or hallucinate on rare topics
• Always recommend customers consult a certified mechanic for safety-critical issues