Hi,
When I use chrombpnet for training with the command below:
chrombpnet pipeline -ibam ../01.cdata/AL.chr.sort.bam -d ATAC -g ../03.basenji/b73.chr.fa -c zmays.chr.sizes -p ALpeak_no_blacklist.bed -n AL_output_negatives.bed -fl fold_0.json -b AL_output_auxiliary/bias_model/models/AL_bias.h5 -o AL_output_auxiliary/chrombpnet_model/
I found that when it tries to use modisco for motif finding by
modisco motifs -i AL_output_auxiliary/chrombpnet_model/auxiliary/interpret_subsample/chrombpnet_nobias.profile_scores.h5 -n 50000 -o AL_output_auxiliary/chrombpnet_model/auxiliary/interpret_subsample/modisco_results_profile_scores.h5 -w 500
All the CPU (256 threads; used in some period, not all the time) and GPU (24Gb) in my system were used. I'm wondering how I can adjust the CPUs used in that process?
Hi,
When I use chrombpnet for training with the command below:
chrombpnet pipeline -ibam ../01.cdata/AL.chr.sort.bam -d ATAC -g ../03.basenji/b73.chr.fa -c zmays.chr.sizes -p ALpeak_no_blacklist.bed -n AL_output_negatives.bed -fl fold_0.json -b AL_output_auxiliary/bias_model/models/AL_bias.h5 -o AL_output_auxiliary/chrombpnet_model/I found that when it tries to use
modiscofor motif finding bymodisco motifs -i AL_output_auxiliary/chrombpnet_model/auxiliary/interpret_subsample/chrombpnet_nobias.profile_scores.h5 -n 50000 -o AL_output_auxiliary/chrombpnet_model/auxiliary/interpret_subsample/modisco_results_profile_scores.h5 -w 500All the CPU (256 threads; used in some period, not all the time) and GPU (24Gb) in my system were used. I'm wondering how I can adjust the CPUs used in that process?