diff --git a/docs/source/en/basics/launch_colossalai.md b/docs/source/en/basics/launch_colossalai.md index 8a6028d6c..32748dae1 100644 --- a/docs/source/en/basics/launch_colossalai.md +++ b/docs/source/en/basics/launch_colossalai.md @@ -131,17 +131,18 @@ with one simple command. There are two ways you can launch multi-node jobs. This is suitable when you only have a few nodes. Let's say I have two nodes, namely `host1` and `host2`, I can start multi-node training with the following command. Compared to single-node training, you must specify the `master_addr` -option, which is auto-set to localhost if running on a single node only. +option, which is auto-set to localhost if running on a single node only. \ +Additionally, you must also ensure that all nodes share the same open ssh port, which can be specified using --ssh-port. :::caution -`master_addr` cannot be localhost when running on multiple nodes, it should be the hostname or IP address of a node. +`master_addr` cannot be localhost when running on multiple nodes, it should be the **hostname or IP address** of a node. ::: ```shell # run on these two nodes -colossalai run --nproc_per_node 4 --host host1,host2 --master_addr host1 test.py +colossalai run --nproc_per_node 4 --host host1,host2 --master_addr host1 test.py --ssh-port 22 ``` - Run with `--hostfile` diff --git a/docs/source/zh-Hans/basics/launch_colossalai.md b/docs/source/zh-Hans/basics/launch_colossalai.md index a80d16717..9e40f64c2 100644 --- a/docs/source/zh-Hans/basics/launch_colossalai.md +++ b/docs/source/zh-Hans/basics/launch_colossalai.md @@ -116,17 +116,17 @@ colossalai run --nproc_per_node 4 --master_port 29505 test.py - 通过`--hosts`来启动 这个方式适合节点数不多的情况。假设我们有两个节点,分别为`host`和`host2`。我们可以用以下命令进行多节点训练。 -比起单节点训练,多节点训练需要手动设置`--master_addr` (在单节点训练中`master_addr`默认为`127.0.0.1`)。 +比起单节点训练,多节点训练需要手动设置`--master_addr` (在单节点训练中`master_addr`默认为`127.0.0.1`)。同时,你需要确保每个节点都使用同一个ssh port。可以通过--ssh-port设置。 :::caution -多节点训练时,`master_addr`不能为`localhost`或者`127.0.0.1`,它应该是一个节点的名字或者IP地址。 +多节点训练时,`master_addr`不能为`localhost`或者`127.0.0.1`,它应该是一个节点的**名字或者IP地址**。 ::: ```shell # 在两个节点上训练 -colossalai run --nproc_per_node 4 --host host1,host2 --master_addr host1 test.py +colossalai run --nproc_per_node 4 --host host1,host2 --master_addr host1 test.py --ssh-port 22 ```