Open SingleCellW opened 1 month ago
Hi there, there may be more typos in code, In first several lines in agglomoration_script.py file, dataset_name_list should be changed to the right tasks. While Spectral Clustering, add footprint in the final part should make it work, like "agglomoration.footprint=global_pooling". And there'd better be a mem release in the last part of hierarchical_agglomoration_utils.py, like "plt.cla()", "plt.close("all")" to avoid wasted mem (my 32GB mem v-server may not be suffice).
It's a little painstaking to work through the README part, but still great work, appreciated!
In lotus\skill_learning\skill_discovery\hierarchical_agglomoration.py file Line 49, the judgement in dataset_category has a typo, where 'libero_spactial' should be 'libero_spatial', or demostations in project's README would not work.
Please fix it, thanks.