https://github.com/lukaszbinden/spatial-broadcast-decoder

Spatial Broadcast Decoder Tensorflow Implementation
https://github.com/lukaszbinden/spatial-broadcast-decoder

deep-neural-networks representation-learning tensorflow

Last synced: over 1 year ago
JSON representation

Spatial Broadcast Decoder Tensorflow Implementation

• Host: GitHub
• URL: https://github.com/lukaszbinden/spatial-broadcast-decoder
• Owner: lukaszbinden
• License: mit
• Created: 2019-03-09T18:41:35.000Z (over 7 years ago)
• Default Branch: master
• Last Pushed: 2019-03-11T09:44:12.000Z (over 7 years ago)
• Last Synced: 2025-02-26T20:22:26.021Z (over 1 year ago)
• Topics: deep-neural-networks, representation-learning, tensorflow
• Language: Python
• Size: 9.77 KB
• Stars: 6
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# Spatial Broadcast Decoder

Spatial Broadcast Decoder Implementation using Tensorflow v1.12. This implements only the decoder as proposed in the paper.

See the original paper for specifications:

**Spatial Broadcast Decoder: A Simple Architecture for Learning Disentangled Representations in VAEs**

https://arxiv.org/abs/1901.07017

Note that the convolutional part of the spatial broadcast decoder has a different implementation than in the paper. Just go ahead and adjust the channel numbers for each layer. Kernel and stride are original.