Genesis icon indicating copy to clipboard operation
Genesis copied to clipboard

when building a docker image, it always fails at RUN chmod +x ./build_luisa.sh && ./build_luisa.sh ${PYTHON_VERSION}

Open zhrli opened this issue 10 months ago • 14 comments
trafficstars

I can not take a snapshot.

Anyone meets the same situation?

zhrli avatar Dec 26 '24 08:12 zhrli

Can you provide me your system information (hardware specs, cuda version, etc.) and the exact error message that you are getting?

Kashu7100 avatar Dec 26 '24 08:12 Kashu7100

Can you provide me your system information (hardware specs, cuda version, etc.) and the exact error message that you are getting?

Can you provide me your system information (hardware specs, cuda version, etc.) and the exact error message that you are getting? CPU: 72 在线 CPU 列表: 0-71 厂商 ID: GenuineIntel 型号名称: Intel(R) Xeon(R) Platinum 8352V CPU @ 2.10GHz CPU 系列: 6 型号: 106 每个核的线程数: 2

GPU Info (if applicable): Thu Dec 26 16:37:49 2024 +---------------------------------------------------------------------------------------+ | NVIDIA-SMI 535.113.01 Driver Version: 535.113.01 CUDA Version: 12.2 | |-----------------------------------------+----------------------+----------------------+ | GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. | | | | MIG M. |

4090 两张

Operating System: Linux ubuntu 6.2.0-26-generic https://github.com/Genesis-Embodied-AI/Genesis/issues/26~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Thu Jul 13 16:27:29 UTC 2 x86_64 x86_64 x86_64 GNU/Linux

CUDA Version: nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2023 NVIDIA Corporation Built on Tue_Jun_13_19:16:58_PDT_2023 Cuda compilation tools, release 12.2, V12.2.91 Build cuda_12.2.r12.2/compiler.32965470_0

Python and Libraries: Python 3.11.7

When building luisa , my computer just collapsed. So, I don't have any info about it

zhrli avatar Dec 26 '24 08:12 zhrli

Can you tell me which line of build_luisa.sh causing you the failure at least? With the current information provided, I have no clue what is happening.

Kashu7100 avatar Dec 26 '24 09:12 Kashu7100

When building luisa , my computer just collapsed. So, I don't have any info about it

你这段过了么,如果你在国内的话我怀疑是git没整完超时了 git submodule update --init --recursive

Shikairan avatar Dec 26 '24 09:12 Shikairan

When building luisa , my computer just collapsed. So, I don't have any info about it

你这段过了么,如果你在国内的话我怀疑是git没整完超时了 git submodule update --init --recursive

过了,是超时了好几次。

zhrli avatar Dec 26 '24 09:12 zhrli

Can you tell me which line of build_luisa.sh causing you the failure at least? With the current information provided, I have no clue what is happening.

I just used a MAC to reproduce that. CPU load is 100% full.

1735205241539

zhrli avatar Dec 26 '24 09:12 zhrli

Can you tell me which line of build_luisa.sh causing you the failure at least? With the current information provided, I have no clue what is happening.

and the command I use: docker buildx build --platform linux/amd64 -t genesis -f docker/Dockerfile docker --push

zhrli avatar Dec 26 '24 09:12 zhrli

What happens if you do not do the cross compile (without using buildx)?

Kashu7100 avatar Dec 26 '24 09:12 Kashu7100

What happens if you do not do the cross compile (without using buildx)?

1735210703773

zhrli avatar Dec 26 '24 10:12 zhrli

I also had this error message trying to build the docker on Windows 11. I created a pull request with the necessary changes, after making these I was successfully able to create Genesis! The issue was caused by the script having Windows-style line endings, which Linux couldn't understand. To fix this, I installed a tool to convert those line endings to a Linux-friendly format and made sure the script had the correct permissions to run. I also added a check to confirm the script was in the right place inside the Docker container before running it.

skurtyyskirts avatar Dec 26 '24 11:12 skurtyyskirts

我在 Windows 11 上尝试构建 docker 时也遇到了此错误消息。我创建了一个包含必要更改的拉取请求,完成这些更改后,我成功创建了 Genesis!问题是由脚本具有 Windows 样式的行尾引起的,而 Linux 无法理解。为了解决这个问题,我安装了一个工具将这些行尾转换为 Linux 友好的格式,并确保脚本具有正确的运行权限。我还添加了一个检查,以确认脚本在运行之前位于 Docker 容器内的正确位置。

db70cc11fe793cc9c0d116f925b93fd 微信图片_20241226214315

ningwenjie avatar Dec 26 '24 13:12 ningwenjie

我在 Windows 11 上尝试构建 docker 时也遇到了此错误消息。我创建了一个包含必要更改的拉取请求,完成这些更改后,我成功创建了 Genesis!问题是由脚本具有 Windows 样式的行尾引起的,而 Linux 无法理解。为了解决这个问题,我安装了一个工具将这些行尾转换为 Linux 友好的格式,并确保脚本具有正确的运行权限。我还添加了一个检查,以确认脚本在运行之前位于 Docker 容器内的正确位置。

db70cc11fe793cc9c0d116f925b93fd 微信图片_20241226214315

Thank you for this resolution, but Not work for me though. I have tried in both linux and MAC.

zhrli avatar Dec 27 '24 03:12 zhrli

Same issue for me on Mac Apple silicon. I tried with the non-windows line endings as well. Error logs attached logs.txt

JosvanderWesthuizen avatar Dec 28 '24 01:12 JosvanderWesthuizen

Thank you for the error log. It seems the error comes from the incompatibility of the base pytorch image. I think we need to find an appropriate base image for arm based arch...

InvalidBaseImagePlatform: Base image pytorch/pytorch:2.5.1-cuda12.1-cudnn9-devel was pulled with platform "linux/amd64", expected "linux/arm64" for current build

Kashu7100 avatar Dec 28 '24 02:12 Kashu7100

Can you tell me which line of build_luisa.sh causing you the failure at least? With the current information provided, I have no clue what is happening.

I think this problem is related to #380 . According to that solution, it works on linux.

zhrli avatar Dec 30 '24 03:12 zhrli

@skurtyyskirts I used your dockerfile and still can't build, I've been stuck at this stage for a long time. image

AgentOfLyra avatar Dec 30 '24 13:12 AgentOfLyra

Hello. I was also getting a build error on the ERROR [builder 11/11] RUN chmod +x ./build_luisa.sh && ./build_luisa.sh 3.11: 0.356 /bin/sh: 1: ./build_luisa.sh: not found line. A PR may have already been created and the problem may be solved, but I ran the following command based on the conversation in this thread, and it seems to have succeeded in the build in my environment.

  1. git config --global core.autocrlf input (I think git config core.autocrlf false will probably work too)
  2. Delete the local Genesis repository directory and run git clone again. (Because the core.autocrlf option requires a git checkout) I tried this on the main branch (60704fd).

This is something that all Windows users need to do, so in reality it is probably better to control it with the eol in the .gitattributes file. (Unconfirmed) This may be unnecessary information, but I hope it will be helpful. Sorry if it is not related to this case.

rtmnkt avatar Dec 30 '24 14:12 rtmnkt

@skurtyyskirts I used your dockerfile and still can't build, I've been stuck at this stage for a long time. image

#380

zhrli avatar Dec 30 '24 15:12 zhrli