[Feature Request][ExcuTorch]Add device-level failure

[yangw-dev]

add device level failure to indicate failure of a commit

vercel: demo

if base commit has failures:

• failure_repot column: indicate the failure base commit
• for models: redender Failure -> RightSight Value
  if new commit has failure:
  - failure_repot column: indicate the failure on new commit
  - for models: redender leftSide value -> rightside
  if lcommit === rcommit:
• failure_repot column: indicate the failure
• for models: redender Failure

UI Screenshot:

UI demo with tooltip

next step:
add job-level failure

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Updated (UTC)
torchci	✅ Ready (Inspect)	Visit Preview	Apr 8, 2025 10:09pm

[huydhn]

IMO, having a new FAILURE_REPORT column is not worth the vertical space cost because the column supposes to be empty most of the time. Is this possible to include this information in an existing column? For example, in this dashboard, putting a ⚠ into the model name column like ⚠ meta-llama/Llama-3.2-1B-Instruct-QLORA_INT4_EO8 and put the failure report in the title like https://stackoverflow.com/questions/28492167/display-text-under-cursor. What do you think?

[yangw-dev]

IMO, having a new FAILURE_REPORT column is not worth the vertical space cost because the column supposes to be empty most of the time. Is this possible to include this information in an existing column? For example, in this dashboard, putting a ⚠ into the model name column like ⚠ meta-llama/Llama-3.2-1B-Instruct-QLORA_INT4_EO8 and put the failure report in the title like https://stackoverflow.com/questions/28492167/display-text-under-cursor. What do you think?

yah, i think it's fine to remove the Failure_report column, and I just simple add an icon with tooltip