DS4AI-UPB/qwen25-ro-re-lora

Qwen 2.5 32B QLoRA adapter — Romanian/English Relation Classification

Dragoș Mitruț Vasile · Elena-Simona Apostol · Stefan-Adrian Toma · Adrian Paschke · Ciprian-Octavian Truică

QLoRA adapter for Qwen/Qwen2.5-32B-Instruct, fine-tuned for Relation Classification on a Romanian translation of SemEval-2010 Task 8 plus the original English data. Given a sentence with two marked entities (<e1>, <e2>), the model assigns one of ten directional relations.

Results (macro F1-Score, SemEval-2010 Task 8 test set)

Language	F1-Score	Accuracy
English	0.890	0.880
Romanian	0.864	0.853

Under zero-shot prompting this model scores far lower (0.320 English), but QLoRA fine-tuning brings it level with the best models: it is statistically indistinguishable from a fine-tuned XLM-RoBERTa-large (560M) on Romanian (paired bootstrap, p = 0.13), despite being roughly 57x larger.

Training

• Base: Qwen/Qwen2.5-32B-Instruct, loaded in 4-bit
• LoRA: rank 32, alpha 64, dropout 0.05, all attention and MLP projections
• 3 epochs, effective batch size 16, peak LR 2e-4, cosine decay, 5% warmup
• Combined English + Romanian training data (15,871 examples)
• Single NVIDIA A100 40GB, about 3h28min

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel

base = AutoModelForCausalLM.from_pretrained("Qwen/Qwen2.5-32B-Instruct", load_in_4bit=True, device_map="auto")
model = PeftModel.from_pretrained(base, "DS4AI-UPB/qwen25-ro-re-lora")
tok = AutoTokenizer.from_pretrained("DS4AI-UPB/qwen25-ro-re-lora")

Prompt format and the full relation label set are in the run_inference.py script in the code repository.

Limitations

The Romanian training data is machine-translated with automatic post-validation, not a human gold standard. See the paper and the code repository for the translation quality analysis.

Citation

@misc{vasile2026crosslingual,
  title  = {Cross-lingual Relation Extraction with Large Language Models: Zero-Shot, Few-Shot, and Fine-Tuned Evaluation on Romanian},
  author = {Vasile, Drago\c{s}-Mitru\c{t} and Apostol, Elena-Simona and Toma, \c{S}tefan-Adrian and Paschke, Adrian and Truic\u{a}, Ciprian-Octavian},
  year   = {2026},
  note   = {Preprint}
}