Error at the biginning of the training step : centroid_weights.append(torch.tensor(species_genes_scores[sgn]))

[Hweeseon]

Hi! Thank you for your nice tool. I am tryting to trasfer public mouse data's cell annotation to human data. However while the last step of training there happen to be an issue.

I can't figure out how to solve this situation.

Could you give any advices for this circumstances?

Thank you in advance

[Yanay1]

Hello!

You need to generate centroids for this dataset rather than using the frog and zebrafish centroids. You can change the centroids init path argument to something else and it will create a new file there.

[Hweeseon]

I solved the inital problem then another issue appeared.

my adata.X dataframes are array([[0., 0., 0., ..., 0., 0., 0.], [0., 0., 0., ..., 0., 0., 0.], [0., 0., 0., ..., 0., 0., 0.], ..., [0., 0., 0., ..., 0., 0., 0.], [0., 0., 0., ..., 0., 0., 0.], [0., 0., 0., ..., 0., 0., 0.]], dtype=float32)

format. What should I change?

[Hweeseon]

I solved the error by reffering to previous answer. changing the codes in the script by, scores_df = pd.concat((scores_df, pd.DataFrame([mmd_row])), ignore_index=True)

Thank you