Differences

This shows you the differences between two versions of the page.

--- cryolo_reference_example [2019/04/30 18:30]
twagner [Reference setup]
+++ cryolo_reference_example [2019/09/23 09:25]
twagner [Training]
@@ Line 36: / Line 36: @@
 ===== Training =====
-First, you have to warm up the network with "-g 0" I'm selection GPU 0. It takes 2 minutes 30 seconds on our reference setup:
+The training is done with this command:
 <code>
-cryolo_train.py -c config_phosnet.json -w 5 -g 0
+cryolo_train.py -c config_phosnet.json -w 5 -e 5 -g 0
 </code>
-The actual training is done with this command:
+It needes 5 minutes 50 seconds to converge (5 warmup + 10 "normal" epochs). The best validation loss was 0.03042. These numbers might be a little bit different on your case.
-<code>
-cryolo_train.py -c config_phosnet.json -w 0 -e 10 -g 0
-</code>
-It needed 6 minutes 45 seconds to converge (18 epochs). The best validation loss was 0.03316. These numbers might be a little bit different on your case.
 ===== Prediction =====