Predicting genetic ancestry using the gnomad random forest model

I am trying to use the gnomad-generated random forest model to predict genetic ancestry in UKB. I'm having two issues (below) and wondering if anyone has successfully done this or has any suggestions? Thanks!

1. I cannot export the project PC scores from python. The code (below) runs for a long time but nothing actually writes out. 

v3_pcs_ht = hl.experimental.pc_project(
   mt.GT,
   ht.loadings,
   ht.pca_af,
)

v3_pcs_ht.scores.export("scores.tsv")

2. I cannot get the model to finish running. Once again, the code runs but nothing is output and it never finishes. 

ht, model = assign_population_pcs(
   v3_pcs_ht,
   pc_cols=v3_pcs_ht.scores[:1],
   fit=v3_onx_fit,
   min_prob=v3_min_prob,
   apply_model_func=apply_onnx_classification_model,
)