Skip to content

Conversation

@Meeeee6623
Copy link

Updated the README files for MxNet and TensorFlow for Image Classification.
The Quick Start guides were missing a line crucial to data preperation:
wget -qO- https://raw.githubusercontent.com/soumith/imagenetloader.torch/master/valprep.sh | bash
This line seperates the ImageNet data into classes, and is needed for preprocessing with the given script for TensorFlow.

In addition, for MxNet, the num-classes parameter was passed through from the command line, but never updated in models.py.

Meeeee6623 and others added 26 commits August 31, 2021 20:50
Added missing line: wget -qO- https://raw.githubusercontent.com/soumith/imagenetloader.torch/master/valprep.sh | bash in the Data Processing section of the Quick Start guide. This line makes individual folders for all the classes, and was missing here but present in the MxNet README.
Tabbed over the line: wget -qO- https://raw.githubusercontent.com/soumith/imagenetloader.torch/master/valprep.sh | bash under data preparation in the Quick Start guide. It was appearing outside of the markdown text box.
Added line: wget -qO- https://raw.githubusercontent.com/soumith/imagenetloader.torch/master/valprep.sh | bash in markdown block for dataset preparation under Quick Start guide. This line was present in the MxNet readme, but missing in TensorFlow.
Added line: wget -qO- https://raw.githubusercontent.com/soumith/imagenetloader.torch/master/valprep.sh | bash in markdown block for dataset preparation under Quick Start guide. This line was present in the MxNet readme, but missing in TensorFlow, and necessary for processing with the script given in the next line.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

2 participants