About TensorRT input_shapes

[Hong753]

Hello, I've been trying to convert the model to TensorRT in the newest version via torch_tensorrt without success. From your code it seems that the input size for ranks_depth, ranks_bev, ranks_feat is the same (same for interval_starts, interval_lengths), i.e., min_size == opt_size == max_size. However I am getting different sizes depending on camera parameters. I was wondering whether you had similar issues.