提交 b4eede34 编写于 作者: G gongweibao 提交者: GitHub

Merge pull request #234 from jacquesqiao/reduce-fpe

reduce fpe of machine_translation
......@@ -399,7 +399,7 @@ for param in parameters.keys():
```python
optimizer = paddle.optimizer.Adam(
learning_rate=5e-5,
regularization=paddle.optimizer.L2Regularization(rate=1e-3))
regularization=paddle.optimizer.L2Regularization(rate=8e-4))
trainer = paddle.trainer.SGD(cost=cost,
parameters=parameters,
update_equation=optimizer)
......@@ -423,7 +423,7 @@ for param in parameters.keys():
trainer.train(
reader=wmt14_reader,
event_handler=event_handler,
num_passes=10000,
num_passes=2,
feeding=feeding)
```
......
......@@ -361,7 +361,7 @@ for param in parameters.keys():
```python
optimizer = paddle.optimizer.Adam(
learning_rate=5e-5,
regularization=paddle.optimizer.L2Regularization(rate=1e-3))
regularization=paddle.optimizer.L2Regularization(rate=8e-4))
trainer = paddle.trainer.SGD(cost=cost,
parameters=parameters,
update_equation=optimizer)
......@@ -388,7 +388,7 @@ for param in parameters.keys():
trainer.train(
reader=wmt14_reader,
event_handler=event_handler,
num_passes=10000,
num_passes=2,
feeding=feeding)
```
......
......@@ -107,7 +107,7 @@ def main():
# define optimize method and trainer
optimizer = paddle.optimizer.Adam(
learning_rate=5e-5,
regularization=paddle.optimizer.L2Regularization(rate=1e-3))
regularization=paddle.optimizer.L2Regularization(rate=8e-4))
trainer = paddle.trainer.SGD(
cost=cost, parameters=parameters, update_equation=optimizer)
......@@ -137,7 +137,7 @@ def main():
trainer.train(
reader=wmt14_reader,
event_handler=event_handler,
num_passes=10000,
num_passes=2,
feeding=feeding)
......
......@@ -441,7 +441,7 @@ for param in parameters.keys():
```python
optimizer = paddle.optimizer.Adam(
learning_rate=5e-5,
regularization=paddle.optimizer.L2Regularization(rate=1e-3))
regularization=paddle.optimizer.L2Regularization(rate=8e-4))
trainer = paddle.trainer.SGD(cost=cost,
parameters=parameters,
update_equation=optimizer)
......@@ -465,7 +465,7 @@ for param in parameters.keys():
trainer.train(
reader=wmt14_reader,
event_handler=event_handler,
num_passes=10000,
num_passes=2,
feeding=feeding)
```
......
......@@ -403,7 +403,7 @@ for param in parameters.keys():
```python
optimizer = paddle.optimizer.Adam(
learning_rate=5e-5,
regularization=paddle.optimizer.L2Regularization(rate=1e-3))
regularization=paddle.optimizer.L2Regularization(rate=8e-4))
trainer = paddle.trainer.SGD(cost=cost,
parameters=parameters,
update_equation=optimizer)
......@@ -430,7 +430,7 @@ for param in parameters.keys():
trainer.train(
reader=wmt14_reader,
event_handler=event_handler,
num_passes=10000,
num_passes=2,
feeding=feeding)
```
......
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