File tree Expand file tree Collapse file tree 4 files changed +15
-13
lines changed Expand file tree Collapse file tree 4 files changed +15
-13
lines changed Original file line number Diff line number Diff line change 16
16
$ export DEVICES=$( \l s /dev/nvidia* | xargs -I{} echo ' --device {}:{}' )
17
17
$ docker run ${CUDA_SO} ${DEVICES} -it paddlepaddle/paddle:latest-gpu
18
18
19
- 更多关于Docker的安装与使用, 请参考 `PaddlePaddle Docker 文档 <http://www.paddlepaddle.org/doc_cn/build_and_install/install/docker_install .html >`_ 。
19
+ 更多关于Docker的安装与使用, 请参考 `PaddlePaddle Docker 文档 <http://www.paddlepaddle.org/docs/0.11.0/documentation/zh/getstarted/build_and_install/docker_install_cn .html >`_ 。
20
20
21
21
22
22
2. CMake源码编译, 找到的PythonLibs和PythonInterp版本不一致
Original file line number Diff line number Diff line change 1
1
FAQ
2
2
====
3
3
4
+ 本文档对关于PaddlePaddle的一些常见问题提供了解答。如果您的问题未在此处,请您到 `PaddlePaddle社区 <https://github.com/PaddlePaddle/Paddle/issues >`_ 查找答案或直接提 `issue <https://github.com/PaddlePaddle/Paddle/issues/new >`_ ,我们会及时进行回复。
5
+
4
6
.. toctree ::
5
7
:maxdepth: 1
6
8
Original file line number Diff line number Diff line change @@ -148,24 +148,24 @@ Paddle二进制在运行时捕获了浮点数异常,只要出现浮点数异
148
148
149
149
.. code-block :: python
150
150
151
- optimizer = paddle.optimizer.RMSProp(
152
- learning_rate=1e-3,
153
- gradient_clipping_threshold=10.0,
154
- regularization=paddle.optimizer.L2Regularization(rate=8e-4))
151
+ optimizer = paddle.optimizer.RMSProp(
152
+ learning_rate = 1e-3 ,
153
+ gradient_clipping_threshold = 10.0 ,
154
+ regularization = paddle.optimizer.L2Regularization(rate = 8e-4 ))
155
155
156
156
具体可以参考 `nmt_without_attention <https://github.com/PaddlePaddle/models/blob/develop/nmt_without_attention/train.py#L35 >`_ 示例。
157
157
158
158
2. 设置 :code: `error_clipping_threshold ` 参数,示例代码如下:
159
159
160
160
.. code-block :: python
161
161
162
- decoder_inputs = paddle.layer.fc(
163
- act=paddle.activation.Linear(),
164
- size=decoder_size * 3,
165
- bias_attr=False,
166
- input=[context, current_word],
167
- layer_attr=paddle.attr.ExtraLayerAttribute(
168
- error_clipping_threshold=100.0))
162
+ decoder_inputs = paddle.layer.fc(
163
+ act = paddle.activation.Linear(),
164
+ size = decoder_size * 3 ,
165
+ bias_attr = False ,
166
+ input = [context, current_word],
167
+ layer_attr = paddle.attr.ExtraLayerAttribute(
168
+ error_clipping_threshold = 100.0 ))
169
169
170
170
完整代码可以参考示例 `machine translation <https://github.com/PaddlePaddle/book/blob/develop/08.machine_translation/train.py#L66 >`_ 。
171
171
Original file line number Diff line number Diff line change @@ -196,6 +196,6 @@ PaddlePaddle保存的模型参数文件内容由16字节头信息和网络参数
196
196
obj = " process" ,
197
197
args = {" src_dict_path" : src_dict_path})
198
198
199
- 完整源码可参考 `seqToseq <https://github.com/PaddlePaddle/Paddle/tree /develop/demo/seqToseq >`_ 示例。
199
+ 完整源码可参考 `sequence_recurrent <https://github.com/PaddlePaddle/Paddle/blob /develop/paddle/gserver/tests/sequence_recurrent.py >`_ 示例。
200
200
201
201
You can’t perform that action at this time.
0 commit comments