update dataset after fixing/deleting some puzzles

fix another comment

fix unfinished annotation h4 (change to h4h5)

update files to chessbech+puzzles+lc0

rm old train test80 data

Merge branch 'main' of hf.co:datasets/KeithG33/ChessBot-Dataset

update for lc0 train data

update with lc0 data

update dataset with lc0 data

update README and make it 95:5 split. Now ~35M test images

add train/test files and builder code