Question Answering on SQuAD v1.1 v2.0 (test dev)
89.4F1 ScoreAdam
Evaluation Results
| Method | Links | |
|---|---|---|
| AdamModel=Gemma2-2B, Evaluation Protocol=Fine-tuning2026.02 | 89.4 | |
| LoRAModel=Gemma2-2B, Evaluation Protocol=Fine-tuning2026.02 | 86.9 | |
| ZO-MuonModel=Gemma2-2B, Evaluation Protocol=Fine-tuning2026.02 | 78.4 | |
| LOZOModel=Gemma2-2B, Evaluation Protocol=Fine-tuning2026.02 | 77.3 | |
| S-MeZOModel=Gemma2-2B, Evaluation Protocol=Fine-tuning2026.02 | 76.1 | |
| MeZOModel=Gemma2-2B, Evaluation Protocol=Fine-tuning2026.02 | 74.1 | |
| HiZOOModel=Gemma2-2B, Evaluation Protocol=Fine-tuning2026.02 | 73.4 | |
| SubZeroModel=Gemma2-2B, Evaluation Protocol=Fine-tuning2026.02 | 72.1 |