Eosc/NorBERT3 corpus

Workflow

De-duplication: essentially, removing identical paragraphs using SimHash (similar to the NearDup approach in this paper, although they used MinHash; MurMurHash is another option).
Cleaning

We plan to create two versions of the training corpus:

baseline (as is)
Wikipedia+NCC+NAK multiplied by two to match the C4 size (oversampling quality data)

Starting with 50K, following NorBERT-2. May be later experiment with other values.

Q: The size of NBDigital is 662M tokens. Should we use it? It probably overlaps a lot with NCC.

A: No, it isn't. Only 60 paragraphs out of total 18M in NCC are duplicates of paragraphs in NBDigital. Thus, we definitely should use it.

Q: How should we split training corpora: one sentence per line, one paragraph per line, one document per line?

A: BERT assumes that there is one sentence per line.