Abhinav issues

Repositories
Issues
Comments

Results 3 issues of


                                            Abhinav

Fix: Remove redundant epsilon in loss mask weight calculation

This PR fixes incorrect use of epsilon in mask weight calculation that caused numerical instability when all mask values are False. In `keras/src/losses/loss.py`, the mask weight calculation used: ``` valid...

size:S

Update checkout action version from v4 to v5

### The documents referred to the older version of github-actions ### What's being changed: ### Check off the following: - [ ] A subject matter expert (SME) has reviewed the...

content

triage

github_actions

Add Standalone Mixture of Experts example

Add standalone Mixture-of-Experts (MoE) layers for Keras, usable as drop-in replacements for Dense and Conv2D. Includes full example on CIFAR-10 demonstrating: - DenseMoE: soft-routed expert networks for fully connected layers...