未验证 提交 ada82a3e 编写于 作者: Y Yancey 提交者: GitHub

Add is_local paramter description (#8893)

* add is_local paramter description

* update

* update by comment
上级 2cc2fb41
...@@ -71,6 +71,13 @@ paddle.init( ...@@ -71,6 +71,13 @@ paddle.init(
- trainer_id:**必选,默认0**,每个trainer的唯一ID,从0开始的整数 - trainer_id:**必选,默认0**,每个trainer的唯一ID,从0开始的整数
- pservers:**必选,默认127.0.0.1**,当前训练任务启动的pserver的IP列表,多个IP使用“,”隔开 - pservers:**必选,默认127.0.0.1**,当前训练任务启动的pserver的IP列表,多个IP使用“,”隔开
```python
trainer = paddle.trainer.SGD(..., is_local=False)
```
参数说明
- is_local: **必选, 默认True**, 是否使用PServer更新参数
## 准备数据集 ## 准备数据集
......
...@@ -73,6 +73,14 @@ Parameter Description ...@@ -73,6 +73,14 @@ Parameter Description
- trainer_id: **required, default 0**, ID for every trainer, start from 0. - trainer_id: **required, default 0**, ID for every trainer, start from 0.
- pservers: **required, default 127.0.0.1**, list of IPs of parameter servers, separated by ",". - pservers: **required, default 127.0.0.1**, list of IPs of parameter servers, separated by ",".
```python
trainer = paddle.trainer.SGD(..., is_local=False)
```
Parameter Description
- is_local: **required, default True**, whether update parameters by PServer.
## Prepare Training Dataset ## Prepare Training Dataset
Here's some example code [prepare.py](https://github.com/PaddlePaddle/Paddle/tree/develop/doc/howto/usage/cluster/src/word2vec/prepare.py), it will download public `imikolov` dataset and split it into multiple files according to job parallelism(trainers count). Modify `SPLIT_COUNT` at the begining of `prepare.py` to change the count of output files. Here's some example code [prepare.py](https://github.com/PaddlePaddle/Paddle/tree/develop/doc/howto/usage/cluster/src/word2vec/prepare.py), it will download public `imikolov` dataset and split it into multiple files according to job parallelism(trainers count). Modify `SPLIT_COUNT` at the begining of `prepare.py` to change the count of output files.
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册