ctuning / ck-env

CK repository with components and automation actions to enable portable workflows across diverse platforms including Linux, Windows, MacOS and Android. It includes software detection plugins and meta packages (code, data sets, models, scripts, etc) with the possibility of multiple versions to co-exist in a user or system environment.
https://github.com/mlcommons/ck
BSD 3-Clause "New" or "Revised" License
71 stars 25 forks source link

Package not getting installed in vitis ai #118

Open ChrisillaWyns opened 10 months ago

ChrisillaWyns commented 10 months ago

(vitis-ai-tensorflow) Vitis-AI /workspace > python3 -m ck install package:imagenet-2012-val-min

Searching if CK environment for this package already exists using:

CK environment not found for this package ...

*** Installation path used: /home/vitis-ai-user/CK-TOOLS/dataset-imagenet-ilsvrc2012-val-min

Installing to /home/vitis-ai-user/CK-TOOLS/dataset-imagenet-ilsvrc2012-val-min

Checking whether '/home/vitis-ai-user/CK-TOOLS/dataset-imagenet-ilsvrc2012-val-min/ILSVRC2012_img_val.tar' already exists ... /home/vitis-ai-user/CK-TOOLS/dataset-imagenet-ilsvrc2012-val-min/ILSVRC2012_img_val.tar already exists ...

   CK detected a PROBLEM in the third-party CK package:

   CK package:           imagenet-2012-val-min
   CK repo:              ck-env
   CK repo URL:          https://github.com/ctuning/ck-env
   CK package URL:       https://github.com/ctuning/ck-env/tree/master/package/imagenet-2012-val-min
   Issues URL:           https://github.com/ctuning/ck-env/issues

   CK stable package URL: https://cKnowledge.io/c/package/imagenet-2012-val-min

   Please, check that there is no discussion about this issue at https://cKnowledge.io/c/package/imagenet-2012-val-min

   Please, submit the original comand line and the *full* log to:
     * https://github.com/ctuning/ck/issues
     * https://github.com/ctuning/ck-mlops/issues

CK error: [package] package installation failed!

gfursin commented 10 months ago

Hi @ChrisillaWyns,

This CK version was deprecated 2 years ago and substituted with the new version 3 (aka CM) developed by the cTuning foundation and MLCommons Task Force on Automation and Reproducibility.

I suggest you to switch to the new community version. You can download the min ImageNet val following this CM tutorial:

pip install cmind
cmr "get original imagenet dataset _2012-500"

Our Task Force helped most users successfully migrate from outdated CK to CM. Please join our public Discord server to get help with migration.

You can also learn more about the new version of CK in this ACM video and you may be interested to check the tutorials to run MLPerf benchmarks via CM here.

If you need any further help with CK, don't hesitate to ask!

gfursin commented 10 months ago

By the way, can you please tell us which CK tutorial are you using to run these commands? I thought we updated all tutorials with CM?