[Feature] Semantic model report #29
Description
If semantic model is pre-trained: what is the coverage of the current corpus, e.g. in terms of how many unique and total words?
If semantic model is trained on corpus: any cue as to the dimensions identified? How would this model be reported?