9 Supplementary Material

In figure 9.1 we show the Lift-Splat Adapt diagram. We use the config parameters below. We include a video of example translations. We use the same training parameters from the official release of "Lift, Splat, Shoot" We include a video of example predictions for "Lift Splat" and our best "Lift We train for 35 epochs and use a learning rate of 0.01 with polynomial decay (0.70). We use SGD with Nesterov Momentum (0.9). All the other hyperparameters are kept the same from the no adaptation version. We include a video of example predictions for "Point Pillars" and our best "Point We train for 35 epochs and use a learning rate of 0.01 with polynomial decay (0.70) .