QishengL
QishengL
0 because you don't want to use the inner product of itself. 1 means other positions are good to use. After mask * logits_mask. You get all the positions that...
I still did not understand the calculation of loss. Did you figure it out? Can anyone explain a little bit more?
I followed the setup guide in Windows. I retried it in a Linux system. No error this time. Maybe I should retry it in Windows.
I solved this after downgrading Transformer to 4.2x
I think the problem is nothing inside the generator. If I do next to that I get File "/python3.10/site-packages/torch/nn/parallel/scatter_gather.py", line 69, in return type(out)((k, gather_map([d[k] for d in outputs])) TypeError:...
我解决了。I figure it out。。。
Damn I try to use the same way to create second and third wallet but it failed again. This is not the solution. Maybe just try to create new wallets...
sorry I thought I find the solution before… it suddenly worked for one of my new generated wallet after I do a lot of operation. But it failed again so...
> > sorry I thought I find the solution before… it suddenly worked for one of my new generated wallet after I do a lot of operation. But it failed...