gmm

Windows上 GPU paddle跑CPU模型时报WITH_NCCL的错 paddle=1.8.4.post107 python=3.6\2.7 报错模型 fibinet\flen \word2vec python -m paddlerec.run -m models/rank/word2vec/config.yaml 使用gpu 版本的paddle跑CPUmode 的时候会报下面的错，如果修改mode为gpu，则正常 

【benchmark】 add llama-7b_auto_dp2mp2pp2 benchmark script for cinn

2

### PR types Others ### PR changes Others ### Description add llama-7b_auto_dp2mp2pp2 benchmark script for cinn

【Develop】fix alt_gvt_base

Why haven't the performance data for PyTorch BERT been updated since 2022? Is this repo no longer maintained? *Examples:* * *PyTorch* **Additional context** 

依赖版本： torch 2.3.0+cu118 torchaudio 2.3.0+cu118 torchvision 0.18.0+cu118 MMCV==2.2.0 执行命令： ` ./tools/dist_train.sh petrv2_vovnet_gridmask_p4_800x320_dn_centerview_benchmark_configs/petrv2_vovnet_gridmask_p4_800x320_dn_centerview_bs1_fp16_DP.py 1 --work-dir work_dirs/petrv2_vovnet_gridmask_p4_800x320_dn_centerview_bs1_fp16_DP/ ` 报错： Traceback (most recent call last): File "/PETR/./tools/train.py", line 26, in from mmdet import...

【benchmark】add timeout for test_tipc/benchmark_train.sh

1

【benchmark】add timeout for test_tipc/benchmark_train.sh