ys = torch.cat((g_xs, Gs_joint), 1 this line of code give Cuda out of memory error .how to fix it ?

microsoft / Relation-Aware-Global-Attention-Networks

We design an effective Relation-Aware Global Attention (RGA) module for CNNs to globally infer the attention.

MIT License

337 stars 65 forks source link

Open ssbilakeri opened 3 years ago

Francis-Rings commented 3 years ago

Do you have soled the problem yet? I meet the same problem.

hppy152 commented 2 years ago

You can decrease the batchsize in .sh