Base Training error

[Miracle-hpf]

Sorry for troubling you.I follow your instruction and when I run train_meta.py, a runtimeerror occured as follow:

Could you tell me how to solve it?

[eghouti]

you can solve it as follows tmp = buf[start:start+num_w] tmp = tmp.view(conv_model.weight.shape) conv_model.weight.data.copy(tmp)

[Miracle-hpf]

you can solve it as follows tmp = buf[start:start+num_w] tmp = tmp.view(conv_model.weight.shape) conv_model.weight.data.copy(tmp)

Thanks for your reply!I reduce the torch version to 0.3.1 and now can run.

[zhongdajian]

you can solve it as follows tmp = buf[start:start+num_w] tmp = tmp.view(conv_model.weight.shape) conv_model.weight.data.copy(tmp)

Hello, i have some questions about the training samplse, can you help me? 'train = ./voc_train.txt' in the 'cfg/metayolo.data', it contains 20 categories, so it contains the novel categories. Should we use 15 base categories instead?