You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Jan 13, 2022. It is now read-only.
Hello,
I was following the guideline from the website to run the train_flipflop.py and guppy_basecaller code.
However I only have cpu, so how would I revise the template to run the code. Thx in advance for the help!
Hello. For train_flipflop.py, if --device is not given, or --device cpu is used, then training will be performed on the CPU; setting the environment variable OMP_NUM_THREADS to an appropriate value will critical for achieving maximum performance here. We highly recommend using a GPU since the difference in speed is about two orders of magnitude over CPU only.
Queries about Guppy are best addressed to the appropriate forum. I believe a CPU version of guppy is available, and the standard version may be capable of running in CPU only mode.
Sign up for freeto subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Hello,
I was following the guideline from the website to run the train_flipflop.py and guppy_basecaller code.
However I only have cpu, so how would I revise the template to run the code. Thx in advance for the help!
Script:
train_flipflop.py --device 0 taiyaki/models/mLstm_cat_mod_flipflop.py modbase.hdf5
guppy_basecaller --input_path /path/to/input_reads --save_path /path/to/save_dir --config dna_r9.4.1_450bps_flipflop.cfg --model path/to/model.json --device cuda:1
The text was updated successfully, but these errors were encountered: