Open AnnALekhinaE opened 3 years ago
Hey! I trained yolo3 with lstm and spp layers. How can you optimize the code so that your configurator can read custom route85, output and bottleneck layers? If I delete these functions, then the frozen model cannot classify objects at all.
Showing an example of an lstm layer in netron
Hey! I trained yolo3 with lstm and spp layers. How can you optimize the code so that your configurator can read custom route85, output and bottleneck layers? If I delete these functions, then the frozen model cannot classify objects at all.