ml5js/ml5-library

Default Learning Rates for Optimizers are higher than those in Tensorflow.js/Keras

Open

#1,249 opened on Nov 13, 2021

View on GitHub
 (2 comments) (0 reactions) (0 assignees)JavaScript (6,136 stars) (905 forks)batch import
bughelp wantedneural-network

Description

Dear ml5 community,

I'm submitting a new issue. Please see the details below.

→ Step 1: Describe the issue 📝

The default Optimizer learning rates are significantly higher than those in tensorflow.js. This can lead to sub-optimal accuracy rates when training.

→ Step 2: Screenshots or Relevant Documentation 🖼

Default ML5.js Regression Adam Optimizer learning rate: .20 Default Tensorflow.js Adam Optimizer learning rate: .001

Contributor guide

Tech stack
javascript
Domain
machine learning
Issue type
bug
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
2
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
under 1 hour
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
stale
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
clear
Prerequisites
JavaScriptbasic knowledge of neural network optimizers
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
70
Research direction
Change the default learning rate for Adam optimizer from 0.2 to 0.001 in src/NeuralNetwork/index.js line 18 to match TensorFlow.js defaults. Ensure other optimizers have consistent defaults. Update any related tests.