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
@barisbatuhan says: I have implemented a residual block myself in my project and I am having an issue in the residual block. If I only predict by giving an input (no backward pass), the method works without an issue. However, if training is enabled, then I am getting this error:
LoadError: MethodError: no method matching copyto!(::KnetArray{Float32,4}, ::AutoGrad.Result{KnetArray{Float32,4}})
@barisbatuhan says: I have implemented a residual block myself in my project and I am having an issue in the residual block. If I only predict by giving an input (no backward pass), the method works without an issue. However, if training is enabled, then I am getting this error:
The code piece I am getting error is given below:
https://drive.google.com/file/d/1jP-QxsWGvIUHo3FKiSmNzt5PChY4Y_Be/view?usp=sharing
Here is the error log:
https://drive.google.com/file/d/1zwGZQTdFTBXoy_acYT2G-o2Bx57Q6qzn/view?usp=sharing
The workaround of using
+
instead of.+
works, but we should fix.+
as well.The text was updated successfully, but these errors were encountered: