This repositoty consists of code used for training the model used for the Master Thesis "Road condition assessment using deep learning" by Johan Bloch Madsen (s152991) and Jonas Gottlieb Svendsen (s152992).
Notes about the code:
The code was run on the DTU HPC batch system and as such was designed to be called from batch scripts and most take some input. Some parameters in the code are hardcoded in each file and need to changed directly. The same are true for some of the code that have references to specific directories. To use the code these might need to checked are set correctly. Due to the large amount of data available in the simulation dataset the training was done based on number of batches instead of epochs. So the training loop look different wheater the scripts was designed to train the model on the simulation dataset or the other much smaller datasets.
\ No newline at end of file
The code was run on the DTU HPC batch system and as such was designed to be called from batch scripts and most take some input. Some parameters in the code are hardcoded in each file and need to be changed directly. The same are true for some of the code that have references to specific directories. To use the code it might need to be checked that these are set correctly to the training that is done. Due to the large amount of data available in the simulation dataset the training was done based on number of batches instead of epochs. So the training loop look different whether the scripts was designed to train the model on the simulation dataset or the other much smaller datasets.
