提交 e0bab5f6 编写于 作者: T tangwei

add log hint for local cluster training

上级 73d3502f
# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
train:
trainer:
# for cluster training
strategy: "async"
epochs: 10
workspace: "fleetrec.models.rank.dnn"
reader:
batch_size: 2
class: "{workspace}/../criteo_reader.py"
train_data_path: "{workspace}/data/train"
model:
models: "{workspace}/model.py"
hyper_parameters:
sparse_inputs_slots: 27
sparse_feature_number: 1000001
sparse_feature_dim: 9
dense_input_dim: 13
fc_sizes: [512, 256, 128, 32]
learning_rate: 0.001
optimizer: adam
save:
increment:
dirname: "increment"
epoch_interval: 2
save_last: True
inference:
dirname: "inference"
epoch_interval: 4
save_last: True
# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import paddle.fluid as fluid
from fleetrec.core.trainers.transpiler_trainer import TranspileTrainer
from fleetrec.core.utils import envs
class UserDefineTrainer(TranspileTrainer):
def __init__(self, config=None):
TranspileTrainer.__init__(self, config)
print("this is a demo about how to use user define trainer in fleet-rec")
def processor_register(self):
self.regist_context_processor('uninit', self.instance)
self.regist_context_processor('init_pass', self.init)
self.regist_context_processor('train_pass', self.train)
def init(self, context):
self.model.train_net()
optimizer = self.model.optimizer()
optimizer.minimize((self.model.get_cost_op()))
self.fetch_vars = []
self.fetch_alias = []
self.fetch_period = self.model.get_fetch_period()
metrics = self.model.get_metrics()
if metrics:
self.fetch_vars = metrics.values()
self.fetch_alias = metrics.keys()
context['status'] = 'train_pass'
def train(self, context):
# run startup program at once
self._exe.run(fluid.default_startup_program())
dataset = self._get_dataset()
epochs = envs.get_global_env("train.epochs")
for i in range(epochs):
self._exe.train_from_dataset(program=fluid.default_main_program(),
dataset=dataset,
fetch_list=self.fetch_vars,
fetch_info=self.fetch_alias,
print_period=self.fetch_period)
context['is_exit'] = True
...@@ -26,22 +26,20 @@ python setup.py install ...@@ -26,22 +26,20 @@ python setup.py install
## ctr-dnn示例使用 ## ctr-dnn示例使用
目前框架内置了ctr-dnn模型,后续会加入更多模型 目前框架内置了多个模型,简单的命令即可使用内置模型开始单机训练和本地1*1模拟训练
示例代码位于FleetRec/demo/下, 当前支持单机训练和本地1*1模拟训练
### 单机训练 ### 单机训练
```bash ```bash
cd FleetRec cd FleetRec
python -m fleetrec.run \ python -m fleetrec.run \
-m demo/ctr-dnn_train.yaml \ -m fleetrec.models.rank.dnn \
-d cpu \ -d cpu \
-e single -e single
# 使用GPU资源进行训练 # 使用GPU资源进行训练
python -m fleetrec.run \ python -m fleetrec.run \
-m demo/ctr-dnn_train.yaml \ -m fleetrec.models.rank.dnn \
-d gpu \ -d gpu \
-e single -e single
``` ```
...@@ -52,7 +50,7 @@ python -m fleetrec.run \ ...@@ -52,7 +50,7 @@ python -m fleetrec.run \
cd FleetRec cd FleetRec
# 使用CPU资源进行训练 # 使用CPU资源进行训练
python -m fleetrec.run \ python -m fleetrec.run \
-m demo/ctr-dnn_train.yaml \ -m fleetrec.models.rank.dnn \
-d cpu \ -d cpu \
-e local_cluster -e local_cluster
``` ```
...@@ -63,7 +61,7 @@ python -m fleetrec.run \ ...@@ -63,7 +61,7 @@ python -m fleetrec.run \
cd FleetRec cd FleetRec
python -m fleetrec.run \ python -m fleetrec.run \
-m demo/ctr-dnn_train.yaml \ -m fleetrec.models.rank.dnn \
-d cpu \ -d cpu \
-e cluster -e cluster
``` ```
......
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