Language Modeling on Nemotron
14.9PerplexityQwen3-0.6B (Pretrained)
Evaluation results
| Method | Links | |
|---|---|---|
| Qwen3-0.6B (Pretrained)2026.05 | 14.9 | |
| DLR-NetDistillation stage=+Logits distillation (Phase 2)2026.05 | 14.9 | |
| DLR-NetDistillation stage=Module-wise distillation (Phase 1)2026.05 | 21.8 |