pt is a more standard way to store pytorch neural net weights and the ftorch library also loads and saves the weights in pt format, so it may be beneficial to convert the current weights in pickle format to pt format, and update the methods saving and loading weights accordingly.
pt
is a more standard way to store pytorch neural net weights and theftorch
library also loads and saves the weights inpt
format, so it may be beneficial to convert the current weights in pickle format topt
format, and update the methods saving and loading weights accordingly.