Add missing rank argument (pytorch#490)

davidyang19971209 · Jan 14, 2019 · 635e64c · 635e64c
1 parent e0929a4
commit 635e64c
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/imagenet/README.md b/imagenet/README.md
@@ -30,7 +30,7 @@ You should always use the NCCL backend for multi-processing distributed training
 ### Single node, multiple GPUs:
 
 ```bash
-python main.py -a resnet50 --lr 0.01 --dist-url 'tcp://127.0.0.1:FREEPORT' --dist-backend 'nccl' --multiprocessing-distributed --world-size 1 [imagenet-folder with train and val folders]
+python main.py -a resnet50 --lr 0.01 --dist-url 'tcp://127.0.0.1:FREEPORT' --dist-backend 'nccl' --multiprocessing-distributed --world-size 1 --rank 0 [imagenet-folder with train and val folders]
 ```
 
 ### Multiple nodes: