Training on custom data - where to put summary ?

[GraphGrailAi]

I want to implement abstractive training with train_from=model_step_148000.pt but where to put article and it's summary?

Should i process 1-5 steps noted in readme.md (from Option 2: process the data yourself) or not?

In issue #79 it is noted that i can simply put summary after @highlight in raw_data/temp.raw_src but there is no @highlight there. https://github.com/nlpyang/PreSumm/issues/79

[binhna]

I want to implement abstractive training with train_from=model_step_148000.pt but where to put article and it's summary?

Should i process 1-5 steps noted in readme.md (from Option 2: process the data yourself) or not?

In issue #79 it is noted that i can simply put summary after @highlight in raw_data/temp.raw_src but there is no @highlight there. #79

Yes, you need to follow the steps in README. And yes, one line of @highlight followed by one line of summary. Here is an example:

[AyeshaSarwar]

Every @highlight means a separate summary?

For our custom data, should we put @highlight for every line in our summary?

I am a little confused. Please if you could help me a little.