LLM Training on Qwen-3 30B-A3B (64 x B200-8)
4.1Iteration Time (s)Megatron-LM
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Megatron-LMHardware=64 x B200-82025.07 | 4.1 | 20.2 | — | |
| AXLearnHardware=64 x B200-82025.07 | 4.31 | 19.22 | — |
Loading the SOTA2 catalog…
| Method | Links | |||
|---|---|---|---|---|
| Megatron-LMHardware=64 x B200-82025.07 | 4.1 | 20.2 | — | |
| AXLearnHardware=64 x B200-82025.07 | 4.31 | 19.22 | — |