Code Generation on MBPP+ (TPS)
496.19Time Per Sample (TPS)LLaDA2-Mini
Evaluation Results
| Method | Links | |
|---|---|---|
| LLaDA2-MiniHardware=two H100 GPUs, Tensor Parallelism=TP=2, Decoding=single-sample2026.06 | 496.19 | |
| MBD-LLaDA2-MiniHardware=two H100 GPUs, Tensor Parallelism=TP=2, Decoding=single-sample2026.06 | 646.73 | |
| LLaDA2-Mini-DMaxHardware=two H100 GPUs, Tensor Parallelism=TP=2, Decoding=single-sample2026.06 | 754.97 | |
| MBD-LLaDA2-Mini-DMaxHardware=two H100 GPUs, Tensor Parallelism=TP=2, Decoding=single-sample2026.06 | 896.65 |