Stoch-predict-with-Tranformer-LSTM icon indicating copy to clipboard operation
Stoch-predict-with-Tranformer-LSTM copied to clipboard

Published 20 hours ago • verified publisher icon gangweiX

Transformer part code obviosly has bug,

Open hyuan-kamuda opened this issue 2 years ago • 3 comments

if you look at the positionencoding + X , the diemension is different, it can't be added:

\models.py", line 137, in forward return x + self.pe[:x.size(0), :] RuntimeError: The size of tensor a (64) must match the size of tensor b (50) at non-singleton dimension 0

hyuan-kamuda avatar May 16 '22 13:05 hyuan-kamuda

@hyuan-kamuda did you found the solution ?

apoorv28goel avatar Apr 17 '23 16:04 apoorv28goel

@apoorv28goel did you find any solution?

Zhang-gong avatar Jul 03 '23 02:07 Zhang-gong

@Zhang-gong did you find any solution?

wingwwwc avatar Aug 18 '23 11:08 wingwwwc

@wingwwwc did you find any solution?

zenanli020119 avatar Oct 02 '24 04:10 zenanli020119