Views
No views yet
| Parameter | Value |
|---|---|
| Base Model | PersianLLaMA-13B (13 billion parameters) |
| Adaptation Method | LoRA (Low-Rank Adaptation) |
| LoRA Rank (r) | 32 |
| LoRA Alpha | 64 |
| Trainable Parameters | ~67M (0.5% of base model) |
| Target Modules | Query, Key, Value, Output, Gate, Up, Down projections |
| Training Language | Persian/Farsi |
| Domain | Scientific Literature |
pip install transformers peft torch1import torch
2from transformers import AutoModelForCausalLM, AutoTokenizer
3from peft import PeftModel
4
5# Load base model and adapter
6base_model = AutoModelForCausalLM.from_pretrained(
7 "ViraIntelligentDataMining/PersianLLaMA-13B",
8 torch_dtype=torch.bfloat16,
9 device_map="auto"
10)
11model = PeftModel.from_pretrained(base_model, "YOUR_USERNAME/PersianSciQA-LoRA")
12tokenizer = AutoTokenizer.from_pretrained("ViraIntelligentDataMining/PersianLLaMA-13B")
13
14# Generate scientific question
15abstract = "Your Persian scientific abstract here"
16prompt = f"چکیده: {abstract}\nسوال:"
17inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
18
19with torch.no_grad():
20 outputs = model.generate(
21 **inputs,
22 max_new_tokens=50,
23 do_sample=True,
24 temperature=0.7,
25 top_p=0.9,
26 repetition_penalty=1.1,
27 pad_token_id=tokenizer.pad_token_id
28 )
29
30question = tokenizer.decode(outputs[0, inputs.input_ids.shape[1]:], skip_special_tokens=True)
31print(f"Generated Question: {question}")1@misc{persiansciqa-lora-2025,
2 title={PersianSciQA-LoRA: Scientific Question Generation for Persian Literature},
3 author={[Your Name]},
4 year={2025},
5 url={https://huggingface.co/YOUR_USERNAME/PersianSciQA-LoRA},
6 note={LoRA adapter for Persian scientific question generation based on PersianLLaMA-13B}
7}