ml5js/ml5-library

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

Open

#1249 aperta il 13 nov 2021

Vedi su GitHub
 (2 commenti) (0 reazioni) (0 assegnatari)JavaScript (905 fork)batch import
bughelp wantedneural-network

Metriche repository

Star
 (6136 star)
Metriche merge PR
 (Nessuna PR mergiata in 30 g)

Descrizione

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

Guida contributor