Coding on OpenCodeInstruct (val)
0.2382LossMuonp
Evaluation Results
| Method | Links | |
|---|---|---|
| MuonpBase Model=Llama3-1B, Pre-training=AdamW2026.06 | 0.2382 | |
| AdamWBase Model=Llama3-1B, Pre-training=AdamW2026.06 | 0.2398 | |
| MuonBase Model=Llama3-1B, Pre-training=AdamW2026.06 | 0.2423 | |
| MuonpBase Model=Llama3-1B, Pre-training=Muon2026.06 | 0.633 | |
| AdamWBase Model=Llama3-1B, Pre-training=Muon2026.06 | 0.653 | |
| MuonBase Model=Llama3-1B, Pre-training=Muon2026.06 | 0.746 |