From 440c942cd318f56c145ba0eceb98f11173be9708 Mon Sep 17 00:00:00 2001 From: apatsekin Date: Sun, 23 Dec 2018 12:03:17 -0500 Subject: [PATCH] Update README.md (steps= for multi-gpu) steps= remark for multi-gpu training --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index c8f1a1c6..03913cb6 100644 --- a/README.md +++ b/README.md @@ -255,7 +255,7 @@ 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 data/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