Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[lmi][neuron] Add smart defaults to LMI Neuron #2345

Merged

Conversation

tosterberg
Copy link
Contributor

Description

Part 2 of the smart defaults for Neuron. Updates:

  • rolling batch default to auto
  • TP degree heuristically determined based on model config
  • max rolling batch size heuristically determined based on model config

Smart defaults are determined by the maximum concurrency for the total instance, so in instances where we are will to make a large tradeoff for latency we will need to override the max batch size and lower it.

@tosterberg tosterberg force-pushed the neuron-smart-defaults-2 branch 8 times, most recently from e5cc61b to 4831054 Compare August 29, 2024 16:37
@tosterberg tosterberg force-pushed the neuron-smart-defaults-2 branch 2 times, most recently from b1221fd to 6ac4dc5 Compare August 29, 2024 19:15
@tosterberg tosterberg merged commit 7aa6513 into deepjavalibrary:master Aug 29, 2024
9 checks passed
@tosterberg tosterberg deleted the neuron-smart-defaults-2 branch August 29, 2024 21:14
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.

2 participants