search

TimoSaemann / caffe-segnet-cudnn5

This repository was a fork of BVLC/caffe and includes the upsample, bn, dense_image_data and softmax_with_loss (with class weighting) layers of caffe-segnet (https://github.com/alexgkendall/caffe-segnet) to run SegNet with cuDNN version 5.
Other
176 stars 127 forks source link

transform_param fails #24

Closed vsuryamurthy closed 6 years ago

vsuryamurthy commented 6 years ago

I am finetuning a pre-trained model (initially trained on cityscapes dataset) using my own dataset. When I incorporate "mirror:true" in transform_param{} in the net.prototxt, I get the following error: error_1 @TimoSaemann Does the dense_image_data_layer support transform_param?

vsuryamurthy commented 6 years ago

I realised that we can crop or mirror using the dense_image_data_layer. Sorry for my mistake.