You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
However, I've tried to run the code with the Hebbian update rule, and I run across several exceptions.
It seems that in pytorch_hebbian/handlers/tensorboard_logger.py in line 8, BaseWeightsHandler should be imported instead of BaseWeightsHistHandler.
Running HebbsRule from hebb.py results with RuntimeError: 1D tensors expected, but got 2D and 1D tensors. I believe it is due to the torch.dot(w, x) operation: in PyTorch, dot only supports computing the dot product of two 1D tensors with the same number of elements, in contrast to NumPy.
Is there a way to fix it? Do you have any plans on continuing working on this project?
The text was updated successfully, but these errors were encountered:
Hi!
I find this repository very curious, thank you!
However, I've tried to run the code with the Hebbian update rule, and I run across several exceptions.
It seems that in
pytorch_hebbian/handlers/tensorboard_logger.py
in line 8,BaseWeightsHandler
should be imported instead ofBaseWeightsHistHandler
.Running HebbsRule from
hebb.py
results withRuntimeError: 1D tensors expected, but got 2D and 1D tensors
. I believe it is due to thetorch.dot(w, x)
operation: in PyTorch,dot
only supports computing the dot product of two 1D tensors with the same number of elements, in contrast to NumPy.Is there a way to fix it? Do you have any plans on continuing working on this project?
The text was updated successfully, but these errors were encountered: