Skip to content

Energy Efficiency #16

Description

@rafecchang

Objective:

  • Reduce model size

Steps:

  1. Train a disterbert model on the optimized hyperparameters. Use this as a baseline model.
  2. Apply purning (likely hyperparameter optimization for only the purning-related hyperparameters.)
  3. Apply post-training quantization (likely hyperparameter optimization for only the quantization-related hyperparameters.)
  4. Fine tune with hyperparameter optimization

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

No labels
No labels

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions