Self-driving car project inspired by Udacity and Nvidia
Drive manually:
python capture2.py -out_dir stream/my_dir
python drive5.py -st_dir stream/my_dir -multi ...
Drive with model:
python capture2.py -out_dir stream/my_dir
python drive5.py -st_dir stream/my_dir -multi -model models/my_model.h5 ...
Interpolate data:
python script_multi rec_data rec_data my_data
Check for duplicates:
python img_similiarity_detection_multi.py -cutoff 0.2 data_sets/my_data
you may set your own cutoff
Augment image set:
python img_augment_tmp.py my_data
Merge data_sets:
python manage_data_multi.py -set my_data -out_set to_train old_set
Train model on image set:
python train4.py -multi -batch 50 -epoch 10 to_train 100
user should specify batch size, num of training epochs and num training steps