search

Xilinx / Vitis-AI

Vitis AI is Xilinx’s development stack for AI inference on Xilinx hardware platforms, including both edge devices and Alveo cards.
https://www.xilinx.com/ai
Apache License 2.0
1.46k stars 630 forks source link

Error while building gpu docker - manifest not found, no image vitis-ai-gpu #960

Closed Werterror closed 2 years ago

Werterror commented 2 years ago

Hi all, I get error while building gpu docker, CUDA version: (base) hyperworks2@hyperworks2-MS-7B17:~$ nvcc --version nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2019 NVIDIA Corporation Built on Sun_Jul_28_19:07:16_PDT_2019 Cuda compilation tools, release 10.1, V10.1.243 (base) hyperworks2@hyperworks2-MS-7B17:~$ /usr/local/cuda/bin/nvcc --version nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2021 NVIDIA Corporation Built on Fri_Dec_17_18:16:03_PST_2021 Cuda compilation tools, release 11.6, V11.6.55 Build cuda_11.6.r11.6/compiler.30794723_0 (base) hyperworks2@hyperworks2-MS-7B17:~$

Return of sudo ./docker_build_gpu.sh: `(base) hyperworks2@hyperworks2-MS-7B17:~/Vitis-AI/docker$ sudo ./docker_build_gpu.sh NOTICE: BY INVOKING THIS SCRIPT AND USING THE SOFTWARE INSTALLED BY THE SCRIPT, YOU AGREE ON BEHALF OF YOURSELF AND YOUR EMPLOYER (IF APPLICABLE) TO BE BOUND TO THE LICENSE AGREEMENTS APPLICABLE TO THE SOFTWARE THAT YOU INSTALL BY RUNNING THE SCRIPT.

BY ELECTING TO CONTINUE, YOU WILL CAUSE THIS SCRIPT FILE TO AUTOMATICALLY INSTALL A VARIETY OF SOFTWARE COPYRIGHTED BY XILINX AND THIRD PARTIES THAT IS SUBJECT TO VARIOUS LICENSE AGREEMENTS THAT APPEAR UPON INSTALLATION, ACCEPTANCE AND/OR ACTIVATION OF THE SOFTWARE AND/OR ARE CONTAINED OR DESCRIBED IN THE CORRESPONDING RELEASE NOTES OR OTHER DOCUMENTATION OR HEADER OR SOURCE FILES. XILINX DOES NOT GRANT TO LICENSEE ANY RIGHTS OR LICENSES TO SUCH THIRD-PARTY SOFTWARE. LICENSEE AGREES TO CAREFULLY REVIEW AND ABIDE BY THE TERMS AND CONDITIONS OF SUCH LICENSE AGREEMENTS TO THE EXTENT THAT THEY GOVERN SUCH SOFTWARE.

BY ELECTING TO CONTINUE, YOU WILL CAUSE THE FOLLOWING SOFTWARE TO BE DOWNLOADED AND INSTALLED ON YOUR SYSTEM. BY ELECTING TO CONTINUE, YOU UNDERSTAND THAT THE INSTALLATION OF THE SOFTWARE LISTED BELOW MAY ALSO RESULT IN THE INSTALLATION ON YOUR SYSTEM OF ADDITIONAL SOFTWARE NOT LISTED BELOW IN ORDER TO OPERATE (SUCH SOFTWARE IS HEREAFTER REFERRED TO AS ‘DEPENDENCIES’) XILINX DOES NOT GRANT TO LICENSEE ANY RIGHTS OR LICENSES TO SUCH DEPENDENCIES LICENSEE AGREES TO CAREFULLY REVIEW AND ABIDE BY THE TERMS AND CONDITIONS OF ANY LICENSE AGREEMENTS TO THE EXTENT THAT THEY GOVERN SUCH DEPENDENCIES

BY ELECTING TO CONTINUE, YOU WILL CAUSE THE FOLLOWING SOFTWARE PACKAGES (AND THEIR RESPECTIVE DEPENDENCIES, IF APPLICABLE) TO BE DOWNLOADED FROM UBUNTU'S MAIN REPO AND INSTALLED ON YOUR SYSTEM: http://us.archive.ubuntu.com/ubuntu/dists/bionic/ http://us.archive.ubuntu.com/ubuntu/dists/bionic/

  1. sudo
  2. git
  3. zstd
  4. tree
  5. vim
  6. wget
  7. bzip2
  8. ca-certificates
  9. curl
  10. unzip
  11. python3-minimal
  12. python3-opencv
  13. python3-venv
  14. python3-pip
  15. python3-setuptools
  16. g++
  17. make
  18. cmake
  19. build-essential
  20. autoconf
  21. libgoogle-glog-dev
  22. libgflags-dev
  23. libunwind-dev
  24. libtool
  25. libgtk2.0-dev
  26. libavcodec-dev
  27. libavformat-dev
  28. libavdevice-dev

BY ELECTING TO CONTINUE, YOU WILL CAUSE THE FOLLOWING SOFTWARE PACKAGES (AND THEIR RESPECTIVE DEPENDENCIES, IF APPLICABLE) TO BE DOWNLOADED FROM ANACONDA REPO AND INSTALLED ON YOUR SYSTEM: https://anaconda.org”

  1. absl-py
  2. astor
  3. attrs
  4. backcall
  5. backports
  6. backports.weakref
  7. blas
  8. bleach
  9. boost
  10. bzip2
  11. ca-certificates
  12. cairo
  13. c-ares
  14. certifi
  15. cffi
  16. chardet
  17. cloudpickle
  18. conda
  19. conda-package-handling
  20. cryptography
  21. cycler
  22. cytoolz
  23. dask-core
  24. dbus
  25. decorator
  26. defusedxml
  27. dill
  28. dpuv1_compiler
  29. dpuv1-rt
  30. dpuv1-rt-ext
  31. dpuv1-rt-neptune
  32. entrypoints
  33. expat
  34. ffmpeg
  35. fontconfig
  36. freeglut
  37. freetype
  38. fribidi
  39. gast
  40. gettext
  41. gflags
  42. giflib
  43. glib
  44. glog
  45. gmp
  46. gnutls
  47. google-pasta
  48. graphite2
  49. graphviz
  50. grpcio
  51. gst-plugins-base
  52. gstreamer
  53. h5py
  54. harfbuzz
  55. hdf5
  56. icu
  57. idna
  58. imageio
  59. importlib_metadata
  60. importlib-metadata
  61. intel-openmp
  62. ipykernel
  63. ipython
  64. ipython_genutils
  65. ipywidgets
  66. jasper
  67. jedi
  68. jinja2
  69. joblib
  70. jpeg
  71. json-c
  72. jsoncpp
  73. jsonschema
  74. jupyter
  75. jupyter_client
  76. jupyter_console
  77. jupyter_core
  78. keras
  79. keras-applications
  80. keras-base
  81. keras-preprocessing
  82. kiwisolver
  83. krb5
  84. lame
  85. ld_impl_linux-64
  86. leveldb
  87. libblas
  88. libboost
  89. libcblas
  90. libedit
  91. libffi
  92. _libgcc_mutex
  93. libgcc-ng
  94. libgfortran-ng
  95. libglu
  96. libiconv
  97. liblapack
  98. liblapacke
  99. libopenblas
  100. libopencv
  101. libopus
  102. libpng
  103. libprotobuf
  104. libsodium
  105. libssh2
  106. libstdcxx-ng
  107. libtiff
  108. libtool
  109. libuuid
  110. libvpx
  111. libwebp
  112. libxcb
  113. libxml2
  114. lmdb
  115. lz4-c
  116. markdown
  117. markupsafe
  118. marshmallow
  119. matplotlib
  120. matplotlib-base
  121. mistune
  122. mkl
  123. mkl_fft
  124. mkl_random
  125. mkl-service
  126. mock
  127. more-itertools
  128. nbconvert
  129. nbformat
  130. ncurses
  131. nettle
  132. networkx
  133. notebook
  134. numpy
  135. numpy-base
  136. olefile
  137. openblas
  138. opencv
  139. openh264
  140. openssl
  141. opt_einsum
  142. packaging
  143. pandas
  144. pandoc
  145. pandocfilters
  146. pango
  147. parso
  148. pexpect
  149. pickleshare
  150. pillow
  151. pip
  152. pixman
  153. pluggy
  154. progressbar2
  155. prometheus_client
  156. prompt_toolkit
  157. prompt-toolkit
  158. protobuf
  159. ptyprocess
  160. py
  161. pybind11
  162. py-boost
  163. pycosat
  164. pycosat
  165. pycparser
  166. pydot
  167. pygments
  168. py-opencv
  169. pyopenssl
  170. pyparsing
  171. pyqt
  172. pyrsistent
  173. pysocks
  174. pytest
  175. pytest-runner
  176. python
  177. python-dateutil
  178. python-gflags
  179. python-graphviz
  180. python-leveldb
  181. python-utils
  182. pytz
  183. pywavelets
  184. pyyaml
  185. pyzmq
  186. qt
  187. qtconsole
  188. qtpy
  189. readline
  190. requests
  191. ruamel_yaml
  192. scikit-image
  193. scikit-learn
  194. scipy
  195. send2trash
  196. setuptools
  197. sip
  198. six
  199. snappy
  200. sqlite
  201. tensorboard
  202. tensorflow
  203. tensorflow-base
  204. tensorflow-estimator
  205. termcolor
  206. terminado
  207. testpath
  208. _tflow_select
  209. threadpoolctl
  210. tk
  211. toolz
  212. tornado
  213. tqdm
  214. traitlets
  215. urllib3
  216. wcwidth
  217. webencodings
  218. werkzeug
  219. wheel
  220. widgetsnbextension
  221. wrapt
  222. x264
  223. xcompiler
  224. xorg-libice
  225. xorg-libsm
  226. xorg-libx11
  227. xorg-libxext
  228. xorg-libxpm
  229. xorg-libxrender
  230. xorg-libxt
  231. xorg-renderproto
  232. xorg-xextproto
  233. xorg-xproto
  234. xz
  235. yaml
  236. yaml-cpp
  237. zeromq
  238. zipp
  239. zlib
  240. zstd

BY ELECTING TO CONTINUE, YOU ACKNOWLEDGE AND AGREE, FOR YOURSELF AND ON BEHALF OF YOUR EMPLOYER (IF APPLICABLE), THAT XILINX IS NOT DISTRIBUTING TO YOU IN THIS FILE ANY OF THE AFORMENTIONED SOFTWARE OR DEPENDENCIES, AND THAT YOU ARE SOLELY RESPONSIBLE FOR THE INSTALLATION OF SUCH SOFTWARE AND DEPENDENCIES ON YOUR SYSTEM AND FOR CAREFULLY REVIEWING AND ABIDING BY THE TERMS AND CONDITIONS OF ANY LICENSE AGREEMENTS TO THE EXTENT THAT THEY GOVERN SUCH SOFTWARE AND DEPENDENCIES

Press any key to continue...

Do you agree to the terms and wish to proceed [y/n]?

Do you agree to the terms and wish to proceed [y/n]? y Sending build context to Docker daemon 86.02kB Step 1/61 : FROM nvidia/cuda:11.0-cudnn8-runtime-ubuntu18.04 manifest for nvidia/cuda:11.0-cudnn8-runtime-ubuntu18.04 not found: manifest unknown: manifest unknown Error response from daemon: No such image: xilinx/vitis-ai-gpu:2.5.0.1260 Error response from daemon: No such image: xilinx/vitis-ai-gpu:2.5.0.1260 Error response from daemon: No such image: xilinx/vitis-ai-gpu:2.5.0.1260 `

janifer112x commented 2 years ago

Hi @Werterror , this error is due to Nvidia updated their docker tag in docker hub, the base image we are using does not exist any more, we just updated the version from 11.0 to 11.0.3, Could you please pull the latest commit from master branch and try again?

Prithvi-Velicheti commented 2 years ago

janifer can you please elaborate the answer , I got the same answer.

Prithvi-Velicheti commented 2 years ago

thanks, seems to be resolved by changing vitis-ai-gpu.Dockerfile

Werterror commented 2 years ago

Hi @Werterror , this error is due to Nvidia updated their docker tag in docker hub, the base image we are using does not exist any more, we just updated the version from 11.0 to 11.0.3, Could you please pull the latest commit from master branch and try again?

It does build now without any errors but when I try to run it it returns error: docker: Error response from daemon: could not select device driver "" with capabilities: [[gpu]]. I didnt find any other Vits-AI related solution then to instal nvidia container-toolkit, but that solution wasnt confirmed. Is there a known solution? I have CUDA ready GPU and CUDA 11.6 installed.

Werterror commented 2 years ago

I found another thread confirming that installation of nvidia container-toolkit solves the problem, so I went through with it and gpu docker works just fine. Here's installation guide: https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/install-guide.html Thank you for your help.