That training uses some manually created annotations (e.g.
To make training faster, a semi-supervised pretraining is applied to the first stage of the whole model (similar to Loss is its own Reward: Self-Supervision for Reinforcement Learning, though here only applied once at the start). (While that paper uses only predictions for the next timestep, here predictions for the next T timesteps are generated via an LSTM.) The basic training method follows the standard reinforcement learning approach from the original Atari paper.Īdditionally, a separation of Q-values in V (value) and A (advantage) - as described in Dueling Network Architectures for Deep Reinforcement Learning - is used.įurther, the model tries to predict future states and rewards, similar to the description in Deep Successor Reinforcement Learning. it can not directly set the steering wheel angle). It is trained (mostly) via reinforcement learning and only has access to the buttons W, A, S and D The resulting AI will automatically steer, accelerate and brake.
This repository contains code to train and run a self-driving truck in Euro Truck Simulator 2.
0 kommentar(er)
