In a university project, I compared gradient methods for multiclass logistic regression with my project team.
A simplified view of the work, with client information removed.
Prepare synthetic and MNIST data for the comparison.
Run the gradient methods with the chosen block and step-size rules.
Compare accuracy and runtime and document the findings.
We wanted to compare how different update and step-size choices affected accuracy and runtime.
The repository includes the algorithms, experiment setup, and accuracy and runtime comparisons.
A documented comparison of optimization methods under the same experiment setup.
A comparative implementation and report, available in the public repository.
LET’S WORK ON YOUR GTM
Tell me about your target accounts, current stack, and the step between research and outreach that needs attention.