Medical Text Classification on AbjadNLP public official (test)
39.34Macro-F1AraBERTv2
Evaluation results
| Method | Links | |
|---|---|---|
| AraBERTv2pooling=hybrid pooling, dropout=multi-sample dropout head, training=fine-tuned2026.02 | 39.34 | |
| multilingual-E5-large2026.02 | 38.04 | |
| CamelBert2026.02 | 36.03 | |
| AraBERTv2+Llama 3.3 70Bsetup=two-stage hybrid pipeline, re-ranking=Llama 3.3 70B Instruct2026.02 | 30.35 | |
| Qwen 3 3Bmode=Feature Extraction2026.02 | 12.78 |