Views
No views yet
| Base model | Qwen2.5-7B-Instruct |
| Method | QLoRA (4-bit NF4) |
| Dataset | 41 Vedic astrology conversations (Hindi, Hinglish, English) |
| Epochs | 3 |
| LoRA rank | 16 |
| Hardware | Kaggle T4 GPU |
1from transformers import AutoModelForCausalLM, AutoTokenizer
2from peft import PeftModel
3
4base = AutoModelForCausalLM.from_pretrained("Qwen/Qwen2.5-7B-Instruct")
5model = PeftModel.from_pretrained(base, "your-username/vedaz-qwen2.5-lora")
6tokenizer = AutoTokenizer.from_pretrained("your-username/vedaz-qwen2.5-lora")
7
8messages = [
9 {"role": "system", "content": "You are Vedaz's AI Vedic astrologer."},
10 {"role": "user", "content": "Meri shaadi kab hogi? DOB 10 Aug 1997, 2:30 PM, Jaipur."}
11]
12text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
13inputs = tokenizer(text, return_tensors="pt")
14output = model.generate(**inputs, max_new_tokens=300, temperature=0.7, do_sample=True)
15print(tokenizer.decode(output[0], skip_special_tokens=True))