Training Throughput on GPT-OSS-20B workload
140,900Throughput (tokens/s)PithTrain
Evaluation results
| Method | Links | |
|---|---|---|
| PithTrainHardware=1×8-B200, Parallelism=PP2,DP1,CP1,EP4, SeqLen=8192, Precision=BF162026.05 | 140,900 | |
| Megatron-LMHardware=1×8-B200, Parallelism=PP2,DP1,CP1,EP4, SeqLen=8192, Precision=BF162026.05 | 129,500 |