mirror/darknet
1
0
Fork 0
mirror of https://github.com/pjreddie/darknet.git synced 2023-08-10 21:13:14 +03:00
Code Issues Projects Releases Wiki Activity

Merge pull request #2095 from apatsekin/patch-1

Browse Source 
Update README.md (steps= for multi-gpu)
This commit is contained in:
Alexey
2019-03-18 20:36:13 +03:00
committed by GitHub
parent fd2012475c 440c942cd3
commit c4716dadd4
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
README.md
View File

@ -290,7 +290,7 @@ More information about training by the link: http://pjreddie.com/darknet/yolo/#t
2. Then stop and by using partially-trained model `/backup/yolov3-voc_1000.weights` run training with multigpu (up to 4 GPUs): `darknet.exe detector train cfg/voc.data cfg/yolov3-voc.cfg /backup/yolov3-voc_1000.weights -gpus 0,1,2,3`
Only for small datasets sometimes better to decrease learning rate, for 4 GPUs set `learning_rate = 0.00025` (i.e. learning_rate = 0.001 / GPUs). In this case also increase 4x times `burn_in =` and `max_batches =` in your cfg-file. I.e. use `burn_in = 4000` instead of `1000`.
Only for small datasets sometimes better to decrease learning rate, for 4 GPUs set `learning_rate = 0.00025` (i.e. learning_rate = 0.001 / GPUs). In this case also increase 4x times `burn_in =` and `max_batches =` in your cfg-file. I.e. use `burn_in = 4000` instead of `1000`. Same goes for `steps=` if `policy=steps` is set.
https://groups.google.com/d/msg/darknet/NbJqonJBTSY/Te5PfIpuCAAJ