search

facebookresearch / MUSE

A library for Multilingual Unsupervised or Supervised word Embeddings
Other
3.18k stars 544 forks source link

Continue Training #134

Open imr555 opened 5 years ago

imr555 commented 5 years ago

Is there a way to continue the training after certain epochs?

Like for unsupervised.py I train for 100000 epochs and then save the embeddings in txt files.

Then can I somehow use the txt files or best_mapping.pth file or datasets to continue from the previous mentioned epoch.

Thanks in advance.

glample commented 5 years ago

Hi,

Maybe you are looking for this? https://github.com/facebookresearch/MUSE/blob/master/unsupervised.py#L72-L74