How to save model on each epoch?

[darentsia]

Is it possible to save model during training (like using callbacks on each epoch finish)?

1. And how to enable logging in your model? With verbosity not a lot of information available.
2. And is it possible to train model with 2 or more documents as input at once?

Also:

3. Is it possible to call building a new model from python? Or just with command line?

Thanks in advance.

[mpagli]

1- Only basic information is available through logging. But what would you like to see logged? 2- You need to push the two documents in one file, a simple cat doc1 doc2 > merge_docs would do. 3- no :(, only the commandline is available to start training

[darentsia]

1 .

That's all the information I got from the training phase. I don't really know if it's training right now, or stucked .

2. Thank you, I'll use this tip!
3. Thanks. Maybe in the next release?)

[darentsia]

And again:

Is it possible to save model during training (like using callbacks on each epoch finish)?

))

[darentsia]

Or continue training model on the same corpus? For example, I trained my model during 2 epochs and want to continue training. Is it possible?

[guptaprkhr]

Hi @daridar , we will be adding support for checkpoints soon.