Issues encountered when trying to reproduce some results

[Steven-Palayew]

Hello,

In gene_embedding_examples.ipynb here, shouldn't gpt_gene_name_to_embedding_clean_text be what gets updated in the for loop since this is what we write to pickle file in the end?

Also, would you be able to share code to generate genept-s embedding for Pancreas dataset and evaluate? The result I get when trying to adapt code from aorta_data_analysis.ipynb is quite different.

Thanks in advance!