princeton-vl / CornerNet

BSD 3-Clause "New" or "Revised" License
2.36k stars 475 forks source link

train error #87

Closed njustymk closed 5 years ago

njustymk commented 5 years ago

0%| | 0/500000 [00:00<?, ?it/s]Unexpected end of /proc/mounts line `overlay / overlay rw,relatime,lowerdir=/var/lib/docker/overlay2/l/GXNUA5534L67ACWHRZLHAUPOO4:/var/lib/docker/overlay2/l/QIBCT7PB24D23KPEP4B4N2AZOO:/var/lib/docker/overlay2/l/2GTAFWJULEUB75EDZVWMFVYCZW:/var/lib/docker/overlay2/l/YW3YY3RCWXJFPAZ6N7KK2UTQAD:/var/lib/docker/overlay2/l/RPJ6BIVX6CZJ5UNYVO4YUHTZCI:/var/lib/docker/overlay2/l/BKSONT4ZT55IAW75QTC573FMRX:/var/lib/docker/overlay2/l/LQ6CZCMEYHNOWC46C4AZJNOEG2:/var/lib/docker/overlay2/l/TMERUY6AXOQPFUJGU5I2GI6ABD,upperdir=/var/lib/docker/overlay2/362cf0' /root/anaconda3/envs/CornerNet/lib/python3.6/site-packages/torch/nn/modules/upsampling.py:122: UserWarning: nn.Upsampling is deprecated. Use nn.functional.interpolate instead. warnings.warn("nn.Upsampling is deprecated. Use nn.functional.interpolate instead.") ./train.sh: line 1: 3835 Segmentation fault (core dumped) CUDA_VISIBLE_DEVICES=5,6 python3 train.py CornerNet

Official answer is: This is caused by a bug in the hwloc library (which is used by NVML and other things to detect the system topology) where it cannot correctly parse long lines in /proc/mounts. Container file systems tend to create those long lines on a regular basis due to the long hashes they use in file and directory names. Fortunately this has no impact on the hwloc functionality we actually need.

So while the error messages may look alarming, they are harmless and can be safely ignored.

Thanks, Cliff

2

Posted 12/01/2017 03:28 PM

My question is how to ignore it?can anyone help me ,thank you.