Issue when running prediction using a new trained model

[mikeefriedman]

Hi,

I am trying to run a pre-trained set (chr 22) of 170 individuals but I get the following error. The model was trained with the default configuration.

Launching XGMix in pre-trained mode... Loading model... Loading and processing query file...

• Number of SNPs from model: 2170
• Number of SNPs from file: 424147
• Number of intersecting SNPs: 2170
• Percentage of model SNPs covered by query file: 100.0%
• Found 725 (33.3947%) different reference variants. Adjusting... Inferring ancestry on query data... Phasing individual 1/1Traceback (most recent call last): File "XGMix-master\XGMIX.py", line 365, in main(args, verbose=verbose, run_simulation=run_simulation, founders_ratios=founders_ratios, File "XGMix-master\XGMIX.py", line 285, in main X_query_phased, label_pred_query_window = model.phase(X_query) File "XGMix-master\Utils\XGMix.py", line 343, in phase X_m, X_p, Y_m, Y_p, history, XGFix_tracker = XGFix(X_m, X_p, base_prob=base_prob, smoother=self.smooth, verbose=verbose) File "XGMix-master\XGFix\XGFIX.py", line 197, in XGFix outs = outs.reshape(-1,2,A) ValueError: cannot reshape array of size 24 into shape (2,39)

Thanks.

[weekend37]

Hi mikeefriedman and thanks for letting us know. Before I dig into this, could you try our newly released codebase and see if you get the same error? If so, please leave open an issue in that repo and I'll make sure we fix it asap. Thanks again.