This shows how to do a UMAP plot of the low dimension embedding generated by CCPA.
@MxMstrmn Just have a brief look at the list of files and then merge it. The notebook should have everything you need to make nice UMAP plots, I could already see quite some clustering happening in the seq2seq embedding.
I also did a bit of cleanup:
Removed everything in notebooks/, since we've used non if it
Removed all notebooks in preprocessing that are for different datasets
Cleaned up the simon/ folder by deleting old yaml configs and moving important plotting scripts to my repo (mt_thesis)
Relevant for #82
This shows how to do a UMAP plot of the low dimension embedding generated by CCPA.
@MxMstrmn Just have a brief look at the list of files and then merge it. The notebook should have everything you need to make nice UMAP plots, I could already see quite some clustering happening in the
seq2seq
embedding.I also did a bit of cleanup:
notebooks/
, since we've used non if itpreprocessing
that are for different datasetssimon/
folder by deleting old yaml configs and moving important plotting scripts to my repo (mt_thesis
).py
version of all remaining notebooks