torch-rnn
torch-rnn copied to clipboard
Use of the generated json during checkpoint?
I see that there's the current states and information during that checkpoint iteration but whats the use of it to the program? Can I pass in the json file parameters to resume the training?