使用docker运行pytorch遇到的问题

Unexpected end of /proc/mounts line `overlay / overlay rw,relatime,lowerdir=/var/lib/docker/overlay2/l/WSCZHC3LN2PAKQQNB7KG2QEMPX:/var/lib/docker/overlay2/l/KGNOWOVQR3VUAORTB4UIVQNPRG:/var/lib/docker/overlay2/l/T3VGRNYHJHPCPP2R35UGIMMHQC:/var/lib/docker/overlay2/l/4PRJXLHR2W3GWLMBCW7VPTQEYH:/var/lib/docker/overlay2/l/7DQDNT5GX665EEIWKRQ7GWUZBW:/var/lib/docker/overlay2/l/M2ZY5PRIIU3AODJPHV35REPA43:/var/lib/docker/overlay2/l/6CVRWVPIH6MLTWY4V6X25TDF2T:/var/lib/docker/overlay2/l/O5YA6VF2O2NGQELBOHIEZFSW7M:/var/lib/docker/overlay2/l/UGUEHK63IZQ5Y'
Unexpected end of /proc/mounts line `64FA4NBZJRO6B:/var/lib/docker/overlay2/l/ZUOONDBWRYUUGYERDBL4BQQ6IB:/var/lib/docker/overlay2/l/Q7FY5A4XPJFPOXNSOLJCRQCVSP:/var/lib/docker/overlay2/l/OQZVEHYOV6T3WAE5XJDIM4SHSB:/var/lib/docker/overlay2/l/NPM7QA2NQ434BLRSYU345QSLEW:/var/lib/docker/overlay2/l/BP7WUAQC2HD3GD5EXPA4ZJYLGY:/var/lib/docker/overlay2/l/EJIDBY4KJFU47NIH5FUFNDEKF5:/var/lib/docker/overlay2/l/2N7EWPIQR5LS6S6C5OQ2STOBXJ:/var/lib/docker/overlay2/l/MSLPOBSAZX7T6QAKPQGQENKB3R:/var/lib/docker/overlay2/l/7EG35C4NIM4WOWSKWJ36UAO6KM:/var/lib/do'
Unexpected end of /proc/mounts line `cker/overlay2/l/IDH7SZFIO2M6V5GBNNCJMYUCAP:/var/lib/docker/overlay2/l/DZOGJ6YZX6S2SZBIQDDBD23ZVK:/var/lib/docker/overlay2/l/SKLRHHWZTYMT6YHR7BLKE43KR5:/var/lib/docker/overlay2/l/KUIIHTQTYPJYBJTNL5GOM7RT3Y:/var/lib/docker/overlay2/l/YHAA4ESNFXS37UUVJRQYOIWFBH:/var/lib/docker/overlay2/l/ENAQOLZ4BVJ2UYJPETWI4QSV34:/var/lib/docker/overlay2/l/IDZTFWI5VTTLYVSWI5QVTT35UQ,upperdir=/var/lib/docker/overlay2/0844b567bba7e223ba58bfa7013698995cf3f8784110b59be90685483c7c5d16/diff,workdir=/var/lib/docker/overlay2/0844b56'

这是一个bug，可以忽略https://devtalk.nvidia.com/default/topic/1027077/container-pytorch/-quot-unexpected-end-of-proc-mounts-line-overlay-quot-on-p3-8xlarge/

ERROR: Unexpected bus error encountered in worker. This might be caused by insufficient shared memory (shm).

解决办法：

docker run --runtime=nvidia -e NVIDIA_VISIBLE_DEVICES=0,1,2,3 --shm-size 8G  -it --rm -p 8888:8888 -v /home/stu/LZG/:/LZG lyx00/pytorch04:latest bash

RuntimeError: Failed to export an ONNX attribute, since it's not constant, please try to make things (e.g., kernel size) static if possible

在使用pytorch1.1.0的时候出现上面的错误，但是可以在pytorch1.0.1上面运行https://github.com/pytorch/pytorch/issues/19374

ImportError: libcublas.so.8.0: cannot open shared object file: No such file or directory

刚开始的docker环境是cuda10.1，重新配了个cuda8.0的镜像。

docker中的python环境配置参考https://blog.csdn.net/BruceLee89/article/details/81871174