The Cerebros package is an ultra-precise Neural Architecture Search (NAS) / AutoML that is intended to much more closely mimic biological neurons than conventional neural network architecture strategies.
Can you provide the code or combination of parameters that was involved (if not, its ok)?
Default parameters
Can you copy and paste in any error or unexpected results?
No errors yet: Issue is that there is a narrow margin between the completion time and time limit. It is likely we will experience fails as an artifact of natural variance in execution timing.
Expected behavior
Should have 1.5 st dev between mean execution time and time limit.
Screenshots / logs
113 min run time, but 120 min time limit. I need to up this to 150 min. On a more global level, we may want to track the history of tests and their run times, to calculate mean and st dev.
- name: Phishing email detection with GPT2 embedding
timeout-minutes: 120
Kind of issue: CICD
Environment: Github
Can you provide the code or combination of parameters that was involved (if not, its ok)?
Default parameters
Can you copy and paste in any error or unexpected results?
No errors yet: Issue is that there is a narrow margin between the completion time and time limit. It is likely we will experience fails as an artifact of natural variance in execution timing.
Expected behavior
Should have 1.5 st dev between mean execution time and time limit.
Screenshots / logs
Execution environment: Github Runner: Ubuntu: Latest
Desktop (please complete the following information):
N/A
Additional context
https://github.com/david-thrower/cerebros-core-algorithm-alpha/commit/7530497d153be3190eb4742cef15dff47ad44c71
Suggested Labels (If you don't know, that's ok):