search

PacificBiosciences / ANGEL

Robust Open Reading Frame prediction (ANGLE re-implementation)
Other
16 stars 14 forks source link

A problem with running angel_train.py #17

Closed lijl459 closed 6 years ago

lijl459 commented 6 years ago

Dear Magdoll, Hellow! When I was running angel_train.py to do ANGEL classifier training, I got some trouble here. The command needs more than 400G memory to run, otherwise the memory of our computer is limit to 124G. So, what can I do to reduce the memory in this step? There are 18000 sequences and 28262787 bases in total in the dumb.final.cds file. Thanks a lot!

Magdoll commented 6 years ago

Hi @lijl459 , You could downsample the ANGEL training set. See tutorial here. You could use --use_top=5000 to limit training to, say, the top 5000 sequences.

--LIz

lijl459 commented 6 years ago

@Magdoll Very thanks for your reply. I will try it latter.