Can't find the meta-testing section in the code

[floatSDSDS]

Hi. We are trying to reproduce the results in the paper and we've met some problems.

If we run train.demo.books.t2.sh, the training and result printing section in run_mnet_rand_supports_large_hybrid.py will use all the 23 domains, as listed in workspace.filtered.list. I can't find how args.targetid or args.trgfilelist are used in the code to filter out the target 4 domains. As a result, I can't locate the meta-testing section in the code.

As an attempt, we DIY the meta-testing section ourselves with the separate trainfilelist and trglist. However, we can't get the comparable test results as performed in the paper. We may have flaws or lose the details. So, is it possible to have a full script to run the experiment?

Thanks.