MLPerf Inference v4.1 - AshokBhat/ml GitHub Wiki
BERT numbers
| Config | HW | Core count | Offline | Server |
|---|---|---|---|---|
| 1-node-2S-EMR-PyTorch | INTEL(R) XEON(R) PLATINUM 8592+ | 64 | 1,666.82 | 1,281.58 |
| 1-node-2S-GNR-PyTorch | INTEL(R) XEON(R) 6980P | 128 | 3,024.03 | 2,436.99 |
RESNET-50 numbers
| Config | HW | Core count | Offline | Server |
|---|---|---|---|---|
| 1-node-2S-EMR-PyTorch | INTEL(R) XEON(R) PLATINUM 8592+ | 64 | 25,204.50 | 22,501.80 |
| 1-node-2S-GNR-PyTorch | INTEL(R) XEON(R) 6980P | 128 | 45,617.30 | 39,798.30 |