Load the ,trained model save, in pytorch for chatbots

Question

I ran the code of this tutorial (link) and after a while learning the model was completed and I chatted with the trained model. After exit the program, in the save directory next to the program file, 8 files with a tar extension were created. I guess they are trained model saved files, what should I do to load those files and reuse them?

Berriel Berriel · Accepted Answer · 2021-04-16T14:40:55

In "Run the Model", they describe what you should do. More specifically, there is this code in their snippet:

# Set checkpoint to load from; set to None if starting from scratch
loadFilename = None
checkpoint_iter = 4000
#loadFilename = os.path.join(save_dir, model_name, corpus_name,
#                            '{}-{}_{}'.format(encoder_n_layers, decoder_n_layers, hidden_size),
#                            '{}_checkpoint.tar'.format(checkpoint_iter))

And you basically need to uncomment the #loadFilename and set the checkpoint_iter to the iteration you want. After that, you can skip the training part, because you already ran it, and run the evaluation code again.

Load the ,trained model save, in pytorch for chatbots

1 Answers