Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/cypw/CRU-Net


https://github.com/cypw/CRU-Net

Last synced: 8 days ago
JSON representation

Awesome Lists containing this project

README

        

# Collective Residual Networks
This repository contains the code and trained models of "Sharing Residual Units Through Collective Tensor Factorization in Deep Neural Networks".

## Implementation

### Augmentation

| Method | Settings |
| :------------- | :--------: |
| Random Mirror | True |
| Random Crop | 8% - 100% |
| Aspect Ratio | 3/4 - 4/3 |
| Random HSL | [20,40,50] |

> Note:
> We did not use PCA Lighting and any other advanced augmentation methods.

### Normalization

The augmented input images are substrated by mean RGB = [ 124, 117, 104 ], and then multiplied by 0.0167.

## Results

### ImageNet-1k

**Single crop validation error (center 224x224 crop from resized image with shorter side=256):**

Model|Setting|Model Size|Top-1
:-----|------:|---------:|:---:
CRU-Net-56 @x14|32x4d|98MB|21.9%
CRU-Net-56 @x14|136x1d|98MB|21.7%
CRU-Net-116 @x28x14|32x4d|168MB|20.6%
CRU-Net-116, wider @x28x14|64x4d|318MB|20.3%

We also trained a tiny CRU-Net-56 with less than half the size of ResNet-50.

**Single crop validation error (center 224x224 crop from resized image with shorter side=256):**

Model|Setting|Model Size|Top-1
:-----|------:|---------:|:---:
CRU-Net-56,tiny @x14|32x4d|48MB|22.9%

### Places365-Standard

**10-crop validation accuracy (averaging softmax scores of 10 224x224 crops from resized image with shorter side=256):**

Model|Setting|Model Size|Top-1
:----|------:|---------:|:---:
CRU-Net-116 @x28x14|32x4d|163MB|56.6%

## Trained Models

Model|Setting|Dataset| Link
:----|------:|:------:|:---:
CRU-Net-56,tiny @x14|32x4d|ImageNet-1k|[GoogleDrive](https://goo.gl/oTG7HJ)
CRU-Net-56 @x14|32x4d|ImageNet-1k|[GoogleDrive](https://goo.gl/AD8Bs0)
CRU-Net-56 @x14|136x1d|ImageNet-1k|[GoogleDrive](https://goo.gl/PEY8al)
CRU-Net-116 @x28x14|32x4d|ImageNet-1k|[GoogleDrive](https://goo.gl/OFyIaD)
CRU-Net-116 @x28x14|32x4d|Places365-Standard|[GoogleDrive](https://goo.gl/6gglTz)