Skip to content
/ RCNN Public

an implementation of 'Recurrent Convolutional Neural Network for Object Recognition'

License

Notifications You must be signed in to change notification settings

LixiaoTHU/RCNN

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

RCNN IN Pytorch

Implementation of Recurrent Convolutional Neural Network for Object Recognition.

Requirements

Refer to requirements.txt

Details

Here is the architecture of the model.

model

Refer to the paper below for more details.

Note

the implementation is a little bit different from the original paper.

• Instead of LRN, using BN here and removing dropout layers except the last linear layer.

• As BN layers introduce extra parameters, the feed-forward filter size in layer 1 is modified to 3 x 3.

• For simplification, using Tencorp (there exsits api in pytorch) for test here instead of the original nine crops.

How to train

usage: train.py [-h] [-n K] [-b BATCH_SIZE] [-e EPOCH] [-s SAVE_DIR] [-l TRAINING_LOG]

Train RCNN

optional arguments: -h, --help show this help message and exit -n K the parameter K for RCNN -b BATCH_SIZE the batch size in just one gpu, * GPU_COUNT -e EPOCH the training epoch -s SAVE_DIR the model parameters to be saved -l TRAINING_LOG the logs to be saved

Results on cifar-10

With Data Augmentation

Network No. of Parameters Testing Error (%)
RCNN-96 0.67 M 7.29
RCNN-128 1.19 M 6.76
RCNN-160 1.86 M 6.66

The result improves slightly compared with that in 4.2.2.Maybe the hyper-parameter can be adjusted to get a better model.

Train procedure for RCNN-96

log_96

About

an implementation of 'Recurrent Convolutional Neural Network for Object Recognition'

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages