Create datasets

First, you need to create training and test datasets, run th createDataSets.lua.

Requires:

• csvigo
• torch7
• wget (UNIX standard command tool)
• zip/unzip (UNIX standard command tool)

This will :

• Download training dataset
• Download test dataset
• Download test set labels
• Unzip training dataset and save it to a torch7 compliant format
• Unzip test dataset and labels, join them and save them to a torch7 compliant format
• Erase temporary created directories on demand

Run th createDataSets.lua -help to see all available options

Load the interactive environment

This script loads the main global variables into Torch environment:

• train_set : training set loaded by dataset.lua
• test_set : test set loaded by dataset.lua
• model : cNN model loaded by models/MSmodel.lua
• criterion : a learning criterion load by models/MSmodel.lua
• learning_rate : loaded by train.lua
• batch_size : loaded by train.lua

This programming architecture is modular, you can use your own preprocessing/train/test functions as well as your models, as long as they respect the model/dataset interface described in the corresponding files (dataset.lua, models/MSmodel.lua, ...)

Just run th -i main.lua to load the elements from the different modules and start interactively changing the model parameters, loading an aldready trained model, tweaking the parameters (learning_rate, batch_size, ...) and using the train() and test() functions.

The first time you run th -i main.lua, the data sets will be preprocessed using the code in preprocessing.lua. You will be asked if you want to save the preprocessed data. Once saved, you can skip this step by using th -i main.lua -use_pp_sets.

If you want to use a different model for instance, just use th -i main.lua -model "path/to/the/model.lua". If you want to load an already trained model, use model = torch.load("path/to/model.t7")in Torch.

Run th main.lua -help to see all available options.

GTSRB Challenge

a.k.a German Traffic Sign Recognition Benchmark 🇩🇪 ⛔ 🚳 🚫 ...

Goal

Use Torch to train and evaluate a 2-stage convolutional neural network able to classify German traffic sign images (43 classes):

• fork the repository under your account,
• go to Settings > Features and enable Issues,
• create an issue under your repo describing your approach,
• report your result(s),
• commit your code,
• edit the README with pre-requisites and usage,
• boost accuracy by experimenting the multi-scale architecture,
• compare with the results obtained in matching mode (i.e use the features with a distance-based search).

Paper

Traffic Sign Recognition with Multi-Scale Convolutional Networks, by Yann LeCun et al.

Dataset

Training

http://benchmark.ini.rub.de/Dataset/GTSRB_Final_Training_Images.zip (263 MB)

Testing

http://benchmark.ini.rub.de/Dataset/GTSRB_Final_Test_Images.zip (84 MB) http://benchmark.ini.rub.de/Dataset/GTSRB_Final_Test_GT.zip (98 kB)