Two step training

[Verythai]

Refer to the paper on http://www.statmt.org/wmt19/pdf/54/WMT12.pdf: "We separated the training process into two steps: the first phase for training a generic model, and the second phase to finetune the model. For the first phase, we trained the model with a union dataset that is the concatenation of eSCAPE-NMT-filtered, and the upsampled official training set by copying 20 times. After reaching the convergence point in the first phase, we fine-tuned the model by running the second phase using only the official training set."

and refer to site readme:

`train_src: train_data.tok.srcmt
train_tgt: train_data.tok.pe

valid_src: dev.tok.srcmt
valid_tgt: dev.tok.pe

save_data: prep-data

...`

How can I save the 2nd step training data to "pre-data" in preprocessing? It gets overwritten or simply with an addition on the pre-data?

Thank you.

[goncalomcorreia]

This is not the repository of the paper you referenced, so I'm not familiar with what the authors of that paper did.