new dataset, attack cases, vae, gat, rec, nav.
-
Implement 6 types of attacks and visualize them for verification.
-
Implemented a 4 node attack.
-
Added equation 2 (step2). Fixed parameters of model_all.model_vae_fixed_parameters and implemented scalable model training, the fixed parameters were later added to model_net.py.
-
Simplifies the image semantic attack code.
-
Added red light input to the navigation model.
-
Add the test agent under carla.
-
Added 2 new attacks: repeat frame, fan masking.