hey every one , i have this problem i trained my model for more than 10h and i used torch version 1.0.0 after that i saved the result in a pth file to use it later in another environment , so i moved to the new env i found out that it works with torch 0.4 and when i tried i load that file again it gives me this error ’ Can’t get attribute ‘_rebuild_parameter’ ’ , im looking for a solotion to my problem which is loading the file of the 1.0.0 version to 0.4 version. i don’t want to train another model for other 10h
Loading checkpoints is not forward compatible.
is there a method to make it compatible ??