Open ygren opened 6 years ago
Hi, thank you! The num_classes
parameter refers to the original ENet and has to be fixed at 12, so it can be loaded from the checkpoint file. This layer is eventually replaced, anyway. The number of instances is not something that is represented as a parameter in this concept. If you have more instances you could also think about increasing the feature_dim
. You would have to adapt the visualization for this.
hello, than you for your wonderful work! I try your work to my dataset, but there is a question. my picture is 5~18 lane line per image ,so I draw my lane line in 19 different colors. and you use 6 different colors,the num_classer is 12. but when I change my num_classes like you in 38. There are some mistakes. but when i change it to 12.Training can be done successfully . So how do I set up this parameter? Can you give me some advice? thx