Skip to content

Conversation

@slippedJim
Copy link
Contributor

when build aiter with dockerfile, can not access rocminfo, try to acquire cu_num from env var.

@slippedJim slippedJim requested a review from valarLip July 31, 2025 03:23
@shengnxu shengnxu merged commit 5b09f07 into main Jul 31, 2025
13 checks passed
@shengnxu shengnxu deleted the jim/dev/fix_get_cu_num branch July 31, 2025 06:21
yzhou103 pushed a commit that referenced this pull request Jul 31, 2025
* try to get cu num from env first

* fix type cast

---------

Co-authored-by: Xu, Shengnan <117875955+shengnxu@users.noreply.github.com>
valarLip added a commit that referenced this pull request Aug 5, 2025
* fix multiprocess tuning problem
add post process to drop abnormal data

* clean code

* fix lint error

* Try to get cu num from env first (#739)

* try to get cu num from env first

* fix type cast

---------

Co-authored-by: Xu, Shengnan <117875955+shengnxu@users.noreply.github.com>

* upadte generate data in a8w8 tuning

* generate data per device

* fix lint error

* clean code and fix moe tuning error

* clean code

* fix lint error

* fix lint error

---------

Co-authored-by: valarLip <103567126+valarLip@users.noreply.github.com>
Co-authored-by: slippedJim <jim.guo@amd.com>
Co-authored-by: Xu, Shengnan <117875955+shengnxu@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants