8000
Skip to content

set enable-expert-parallel for qwen3-235b FP run - #1257

Draft
ccrhx4 wants to merge 2 commits into
HabanaAI:aice/v1.20.1from
ccrhx4:correct_fp8_script_for_qwen3_235b
Draft

set enable-expert-parallel for qwen3-235b FP run#1257
ccrhx4 wants to merge 2 commits into
HabanaAI:aice/v1.20.1from
ccrhx4:correct_fp8_script_for_qwen3_235b

Conversation

@ccrhx4
@ccrhx4 ccrhx4 commented May 14, 2025
Copy link
Copy Markdown

add --enable-expert-parallel option to run Qwen3-235B-A22B FP8 on HPU, otherwise INC will crash.

@ccrhx4
ccrhx4 commented May 14, 2025
Copy link
Copy Markdown
Author

@yangulei Please help to review.

@yangulei yangulei left a comment
Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please apply to benchmark_throughput.sh too.

benchmark_throughput.sh
@ccrhx4
ccrhx4 commented May 15, 2025
Copy link
Copy Markdown
Author

Please apply to benchmark_throughput.sh too.

updated. @yangulei Please help to review.

@yangulei yangulei left a comment
Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@ccrhx4
ccrhx4 commented May 15, 2025
Copy link
Copy Markdown
Author

@czhu15 Please help to merge.

@michalkuligowski michalkuligowski left a comment
Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

not updated for >2 mnths, converting to draft

@michalkuligowski
michalkuligowski marked this pull request as draft July 24, 2025 12:12
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.

3 participants

0