ericmsmall

@HalfPhoton Here are the console outputs from dorado -v. Thanks! ### SUP with batchsize 250 ### dorado basecaller [email protected] pod5s/ -v --modified-bases 6mA 5mC_5hmC --batchsize 250 --reference X.fasta --kit-name SQK-NBD114-96...

At first I thought that might be the case as well, but until dorado is started the GPUs are idle and no memory is being used. Like this: ``` +---------------------------------------------------------------------------------------+...

Apologies, I forgot that I had tried this based on other folks posts. Setting this as an environmental variable, or passing directly to pyTorch had no effect.