提交 ea89bf3e 编写于 作者: K Kexin Zhao

Address comments, add c++ inference example and Merge remote branch

......@@ -137,7 +137,7 @@ include(external/openblas) # download, build, install openblas
include(external/mkldnn) # download, build, install mkldnn
include(external/swig) # download, build, install swig
include(external/warpctc) # download, build, install warpctc
include(external/boost) # download, build, install boost
include(external/boost) # download boost
include(external/any) # download libn::any
include(external/eigen) # download eigen3
include(external/pybind11) # download pybind11
......@@ -156,6 +156,7 @@ include(rdma) # set rdma libraries
include(flags) # set paddle compile flags
include(version) # set PADDLE_VERSION
include(coveralls) # set code coverage
include(inference_lib) # add paddle fluid inference libraries
include_directories("${PADDLE_SOURCE_DIR}")
......
#FROM python:2.7.14
FROM nvidia/cuda:8.0-cudnn5-runtime-ubuntu16.04
RUN apt-get update && apt-get install -y python
RUN pip install -U kubernetes opencv-python && apt-get update -y && apt-get install -y iputils-ping libgtk2.0-dev
# NOTE: By default CI built wheel packages turn WITH_DISTRIBUTE=OFF,
# so we must build one with distribute support to install in this image.
RUN pip install paddlepaddle
RUN sh -c 'echo "import paddle.v2 as paddle\npaddle.dataset.cifar.train10()" | python'
RUN pip uninstall -y paddlepaddle
# below lines may change a lot for debugging
ADD https://raw.githubusercontent.com/PaddlePaddle/cloud/develop/docker/paddle_k8s /usr/bin
ADD https://raw.githubusercontent.com/PaddlePaddle/cloud/develop/docker/k8s_tools.py /root
ADD *.whl /
RUN pip install /*.whl && rm -f /*.whl && \
chmod +x /usr/bin/paddle_k8s
ENV LD_LIBRARY_PATH=/usr/local/lib
ADD vgg16_fluid.py vgg16_v2.py /workspace/
# Performance for Distributed vgg16
## Test Result
### Hardware Infomation
- CPU: Intel(R) Xeon(R) CPU E5-2620 v4 @ 2.10GHz
- cpu MHz : 2101.000
- cache size : 20480 KB
### Single Node Single Thread
- PServer Count: 10
- Trainer Count: 20
- Metrics: samples / sec
| Batch Size | 32 | 64 | 128 | 256 |
| -- | -- | -- | -- | -- |
| PaddlePaddle Fluid | 15.44 | 16.32 | 16.74 | 16.79 |
| PaddlePaddle v2 | 15.97 | 17.04 | 17.60 | 17.83 |
| TensorFlow | - | - | - | - |
### Different Batch Size
- PServer Count: 10
- Trainer Count: 20
- Per trainer CPU Core: 1
- Metrics: samples / sec
| Batch Size | 32 | 64 | 128 | 256 |
| -- | -- | -- | -- | -- |
| PaddlePaddle Fluid | 190.20 | 222.15 | 247.40 | 258.18 |
| PaddlePaddle v2 | 170.96 | 233.71 | 256.14 | 329.23 |
| TensorFlow | - | - | - | - |
### Accelerate Rate
- Pserver Count: 20
- Batch Size: 128
- Metrics: samples / sec
| Trainer Count | 20 | 40 | 80 | 100 |
| -- | -- | -- | -- | -- |
| PaddlePaddle Fluid | 263.29 (78.64%) | 518.80 (77.47%) | 836.26 (62.44%) | 1019.29 (60.89%) |
| PaddlePaddle v2 (need more tests) | 326.85 (92.85%) | 534.58 (75.93%) | 853.30 (60.60%) | 1041.99 (59.20%) |
| TensorFlow | - | - | - | - |
### Different Pserver Count
- Trainer Count: 60
- Batch Size: 128
- Metrics: samples/ sec
| PServer Count | 3 | 6 |10 | 20 |
| -- | -- | -- | -- | -- |
| PaddlePaddle Fluid(should fix in next PR) | 589.1 | 592.6 | 656.4 | 655.8 |
| PaddlePaddle v2 | 593.4 | 791.3 | 729.7 | 821.7 |
| TensorFlow | - | - | - | - |
*The performance gap between Fuild and v2 comes from the network interference.*
## Steps to Run the Performance Test
1. You must re-compile PaddlePaddle and enable `-DWITH_DISTRIBUTE` to build PaddlePaddle with distributed support.
1. When the build finishes, copy the output `whl` package located under `build/python/dist` to current directory.
1. Run `docker build -t [image:tag] .` to build the docker image and run `docker push [image:tag]` to push the image to reponsitory so kubernetes can find it.
1. Run `kubectl create -f pserver.yaml && kubectl create -f trainer.yaml` to start the job on your kubernetes cluster (you must configure the `kubectl` client before this step).
1. Run `kubectl get po` to get running pods, and run `kubectl logs [podID]` to fetch the pod log of pservers and trainers.
Check the logs for the distributed training progress and analyze the performance.
## Enable Verbos Logs
Edit `pserver.yaml` and `trainer.yaml` and add an environment variable `GLOG_v=3` and `GLOG_logtostderr=1` to see what happend in detail.
apiVersion: extensions/v1beta1
kind: ReplicaSet
metadata:
name: vgg16job-pserver
spec:
replicas: 10
template:
metadata:
labels:
paddle-job-pserver: vgg16job
spec:
hostNetwork: true
imagePullSecrets:
- name: job-registry-secret
containers:
- name: pserver
image: "registry.baidu.com/paddlepaddle/fluid_benchmark:vgg16"
imagePullPolicy: Always
ports:
- name: jobport-30236
containerPort: 30236
env:
- name: PADDLE_JOB_NAME
value: vgg16job
- name: MKL_NUM_THREADS
value: "1"
- name: TRAINING_ROLE
value: "PSERVER"
- name: TRAINERS
value: "20"
- name: PSERVERS
value: "10"
- name: TOPOLOGY
value: ""
- name: ENTRY
value: "MKL_NUM_THREADS=1 python /workspace/vgg16_fluid.py --local 0"
- name: TRAINER_PACKAGE
value: "/workspace"
- name: PADDLE_INIT_PORT
value: "30236"
- name: PADDLE_INIT_NICS
value: "xgbe0"
- name: PADDLE_INIT_TRAINER_COUNT
value: "1"
- name: PADDLE_INIT_PORTS_NUM
value: "1"
- name: PADDLE_INIT_PORTS_NUM_FOR_SPARSE
value: "1"
- name: PADDLE_INIT_NUM_GRADIENT_SERVERS
value: "20"
- name: PADDLE_INIT_NUM_PASSES
value: "1"
- name: PADDLE_INIT_USE_GPU
value: "0"
- name: LD_LIBRARY_PATH
value: "/usr/local/lib:/usr/local/nvidia/lib64"
- name: NAMESPACE
valueFrom:
fieldRef:
fieldPath: "metadata.namespace"
- name: POD_IP
valueFrom:
fieldRef:
fieldPath: "status.podIP"
command: ["paddle_k8s", "start_fluid"]
resources:
requests:
memory: 10Gi
cpu: 4
limits:
memory: 10Gi
cpu: 4
apiVersion: batch/v1
kind: Job
metadata:
name: vgg16job-trainer
spec:
parallelism: 20
completions: 20
template:
metadata:
labels:
paddle-job: vgg16job
spec:
imagePullSecrets:
- name: job-registry-secret
hostNetwork: true
containers:
- name: trainer
image: "registry.baidu.com/paddlepaddle/fluid_benchmark:vgg16"
imagePullPolicy: Always
command: ["paddle_k8s", "start_fluid"]
env:
- name: PADDLE_JOB_NAME
value: vgg16job
- name: TRAINING_ROLE
value: "TRAINER"
- name: TRAINERS
value: "20"
- name: PSERVERS
value: "10"
- name: TOPOLOGY
value: ""
- name: ENTRY
value: "MKL_NUM_THREADS=1 python /workspace/vgg16_fluid.py --local 0 --batch_size 128"
- name: TRAINER_PACKAGE
value: "/workspace"
- name: PADDLE_INIT_PORT
value: "30236"
- name: PADDLE_INIT_NICS
value: "xgbe0"
- name: PADDLE_INIT_TRAINER_COUNT
value: "1"
- name: PADDLE_INIT_PORTS_NUM
value: "1"
- name: PADDLE_INIT_PORTS_NUM_FOR_SPARSE
value: "1"
- name: PADDLE_INIT_NUM_GRADIENT_SERVERS
value: "20"
- name: PADDLE_INIT_NUM_PASSES
value: "1"
- name: PADDLE_INIT_USE_GPU
value: "0"
- name: LD_LIBRARY_PATH
value: "/usr/local/lib:/usr/local/nvidia/lib64"
- name: NAMESPACE
valueFrom:
fieldRef:
fieldPath: "metadata.namespace"
- name: POD_IP
valueFrom:
fieldRef:
fieldPath: "status.podIP"
resources:
requests:
memory: 40Gi
cpu: 2
limits:
memory: 40Gi
cpu: 2
restartPolicy: Never
apiVersion: extensions/v1beta1
kind: ReplicaSet
metadata:
name: vgg16v2job-pserver
spec:
replicas: 10
template:
metadata:
labels:
paddle-job-pserver: vgg16v2job
spec:
hostNetwork: true
imagePullSecrets:
- name: job-registry-secret
containers:
- name: pserver
image: "registry.baidu.com/paddlepaddle/fluid_benchmark:vgg16"
imagePullPolicy: Always
ports:
- name: jobport-30236
containerPort: 30236
env:
- name: PADDLE_JOB_NAME
value: vgg16v2job
- name: TRAINERS
value: "20"
- name: PSERVERS
value: "10"
- name: TOPOLOGY
value: ""
- name: ENTRY
value: "python train.py"
- name: TRAINER_PACKAGE
value: "/workspace"
- name: PADDLE_INIT_PORT
value: "30236"
- name: PADDLE_INIT_NICS
value: "xgbe0"
- name: PADDLE_INIT_TRAINER_COUNT
value: "1"
- name: PADDLE_INIT_PORTS_NUM
value: "1"
- name: PADDLE_INIT_PORTS_NUM_FOR_SPARSE
value: "1"
- name: PADDLE_INIT_NUM_GRADIENT_SERVERS
value: "20"
- name: PADDLE_INIT_NUM_PASSES
value: "1"
- name: PADDLE_INIT_USE_GPU
value: "0"
- name: LD_LIBRARY_PATH
value: "/usr/local/lib:/usr/local/nvidia/lib64"
- name: NAMESPACE
valueFrom:
fieldRef:
fieldPath: "metadata.namespace"
command: ["paddle_k8s", "start_pserver"]
resources:
requests:
memory: 10Gi
cpu: 4
limits:
memory: 10Gi
cpu: 4
apiVersion: batch/v1
kind: Job
metadata:
name: vgg16v2job-trainer
spec:
parallelism: 20
completions: 20
template:
metadata:
labels:
paddle-job: vgg16v2job
spec:
imagePullSecrets:
- name: job-registry-secret
hostNetwork: true
containers:
- name: trainer
image: "registry.baidu.com/paddlepaddle/fluid_benchmark:vgg16"
imagePullPolicy: Always
command: ["paddle_k8s", "start_trainer", "v2"]
env:
- name: PADDLE_JOB_NAME
value: vgg16v2job
- name: BATCH_SIZE
value: "256"
- name: TRAINERS
value: "20"
- name: PSERVERS
value: "10"
- name: TOPOLOGY
value: ""
- name: ENTRY
value: "cd /workspace && MKL_NUM_THREADS=1 python /workspace/vgg16_v2.py"
- name: TRAINER_PACKAGE
value: "/workspace"
- name: PADDLE_INIT_PORT
value: "30236"
- name: PADDLE_INIT_NICS
value: "xgbe0"
- name: PADDLE_INIT_TRAINER_COUNT
value: "1"
- name: PADDLE_INIT_PORTS_NUM
value: "1"
- name: PADDLE_INIT_PORTS_NUM_FOR_SPARSE
value: "1"
- name: PADDLE_INIT_NUM_GRADIENT_SERVERS
value: "20"
- name: PADDLE_INIT_NUM_PASSES
value: "2"
- name: PADDLE_INIT_USE_GPU
value: "0"
- name: LD_LIBRARY_PATH
value: "/usr/local/lib:/usr/local/nvidia/lib64"
- name: NAMESPACE
valueFrom:
fieldRef:
fieldPath: "metadata.namespace"
resources:
requests:
memory: 40Gi
cpu: 2
limits:
memory: 40Gi
cpu: 2
restartPolicy: Never
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
# 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.
"""VGG16 benchmark in Fluid"""
from __future__ import print_function
import sys
import time
import numpy as np
import paddle.v2 as paddle
import paddle.v2.fluid as fluid
import paddle.v2.fluid.core as core
import paddle.v2.fluid.profiler as profiler
import argparse
import functools
import os
def str2bool(v):
if v.lower() in ('yes', 'true', 't', 'y', '1'):
return True
elif v.lower() in ('no', 'false', 'f', 'n', '0'):
return False
else:
raise argparse.ArgumentTypeError('Boolean value expected.')
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument(
'--batch_size', type=int, default=128, help="Batch size for training.")
parser.add_argument(
'--learning_rate',
type=float,
default=1e-3,
help="Learning rate for training.")
parser.add_argument('--num_passes', type=int, default=50, help="No. of passes.")
parser.add_argument(
'--device',
type=str,
default='CPU',
choices=['CPU', 'GPU'],
help="The device type.")
parser.add_argument('--device_id', type=int, default=0, help="The device id.")
parser.add_argument(
'--data_format',
type=str,
default='NCHW',
choices=['NCHW', 'NHWC'],
help='The data order, now only support NCHW.')
parser.add_argument(
'--data_set',
type=str,
default='cifar10',
choices=['cifar10', 'flowers'],
help='Optional dataset for benchmark.')
parser.add_argument(
'--local',
type=str2bool,
default=True,
help='Whether to run as local mode.')
args = parser.parse_args()
def vgg16_bn_drop(input):
def conv_block(input, num_filter, groups, dropouts):
return fluid.nets.img_conv_group(
input=input,
pool_size=2,
pool_stride=2,
conv_num_filter=[num_filter] * groups,
conv_filter_size=3,
conv_act='relu',
conv_with_batchnorm=True,
conv_batchnorm_drop_rate=dropouts,
pool_type='max')
conv1 = conv_block(input, 64, 2, [0.3, 0])
conv2 = conv_block(conv1, 128, 2, [0.4, 0])
conv3 = conv_block(conv2, 256, 3, [0.4, 0.4, 0])
conv4 = conv_block(conv3, 512, 3, [0.4, 0.4, 0])
conv5 = conv_block(conv4, 512, 3, [0.4, 0.4, 0])
drop = fluid.layers.dropout(x=conv5, dropout_prob=0.5)
fc1 = fluid.layers.fc(input=drop, size=512, act=None)
bn = fluid.layers.batch_norm(input=fc1, act='relu')
drop2 = fluid.layers.dropout(x=bn, dropout_prob=0.5)
fc2 = fluid.layers.fc(input=drop2, size=512, act=None)
return fc2
def main():
if args.data_set == "cifar10":
classdim = 10
if args.data_format == 'NCHW':
data_shape = [3, 32, 32]
else:
data_shape = [32, 32, 3]
else:
classdim = 102
if args.data_format == 'NCHW':
data_shape = [3, 224, 224]
else:
data_shape = [224, 224, 3]
# Input data
images = fluid.layers.data(name='pixel', shape=data_shape, dtype='float32')
label = fluid.layers.data(name='label', shape=[1], dtype='int64')
# Train program
net = vgg16_bn_drop(images)
predict = fluid.layers.fc(input=net, size=classdim, act='softmax')
cost = fluid.layers.cross_entropy(input=predict, label=label)
avg_cost = fluid.layers.mean(x=cost)
# Evaluator
accuracy = fluid.evaluator.Accuracy(input=predict, label=label)
# inference program
inference_program = fluid.default_main_program().clone()
with fluid.program_guard(inference_program):
test_target = accuracy.metrics + accuracy.states
inference_program = fluid.io.get_inference_program(test_target)
# Optimization
optimizer = fluid.optimizer.Adam(learning_rate=args.learning_rate)
optimize_ops, params_grads = optimizer.minimize(avg_cost)
# Initialize executor
place = core.CPUPlace() if args.device == 'CPU' else core.CUDAPlace(
args.device_id)
exe = fluid.Executor(place)
# test
def test(exe):
accuracy.reset(exe)
for batch_id, data in enumerate(test_reader()):
img_data = np.array(map(lambda x: x[0].reshape(data_shape),
data)).astype("float32")
y_data = np.array(map(lambda x: x[1], data)).astype("int64")
y_data = y_data.reshape([-1, 1])
exe.run(inference_program,
feed={"pixel": img_data,
"label": y_data})
return accuracy.eval(exe)
def train_loop(exe, trainer_prog):
iters = 0
ts = time.time()
for pass_id in range(args.num_passes):
# train
start_time = time.time()
num_samples = 0
accuracy.reset(exe)
with profiler.profiler("CPU", 'total') as prof:
for batch_id, data in enumerate(train_reader()):
ts = time.time()
img_data = np.array(
map(lambda x: x[0].reshape(data_shape), data)).astype(
"float32")
y_data = np.array(map(lambda x: x[1], data)).astype("int64")
y_data = y_data.reshape([-1, 1])
loss, acc = exe.run(
trainer_prog,
feed={"pixel": img_data,
"label": y_data},
fetch_list=[avg_cost] + accuracy.metrics)
iters += 1
num_samples += len(data)
print(
"Pass = %d, Iters = %d, Loss = %f, Accuracy = %f, spent %f"
% (pass_id, iters, loss, acc, time.time() - ts)
) # The accuracy is the accumulation of batches, but not the current batch.
pass_elapsed = time.time() - start_time
pass_train_acc = accuracy.eval(exe)
pass_test_acc = test(exe)
print(
"Pass = %d, Training performance = %f imgs/s, Train accuracy = %f, Test accuracy = %f\n"
% (pass_id, num_samples / pass_elapsed, pass_train_acc,
pass_test_acc))
if args.local:
# Parameter initialization
exe.run(fluid.default_startup_program())
# data reader
train_reader = paddle.batch(
paddle.reader.shuffle(
paddle.dataset.cifar.train10() if args.data_set == 'cifar10'
else paddle.dataset.flowers.train(),
buf_size=5120),
batch_size=args.batch_size)
test_reader = paddle.batch(
paddle.dataset.cifar.test10()
if args.data_set == 'cifar10' else paddle.dataset.flowers.test(),
batch_size=args.batch_size)
train_loop(exe, fluid.default_main_program())
else:
pserver_ips = os.getenv("PADDLE_INIT_PSERVERS") # all pserver endpoints
eplist = []
for ip in pserver_ips.split(","):
eplist.append(':'.join([ip, "6174"]))
pserver_endpoints = ",".join(eplist)
print("pserver endpoints: ", pserver_endpoints)
trainers = int(os.getenv("TRAINERS")) # total trainer count
print("trainers total: ", trainers)
current_endpoint = os.getenv(
"POD_IP") + ":6174" # current pserver endpoint
training_role = os.getenv(
"TRAINING_ROLE",
"TRAINER") # get the training role: trainer/pserver
t = fluid.DistributeTranspiler()
t.transpile(
optimize_ops,
params_grads,
pservers=pserver_endpoints,
trainers=trainers)
if training_role == "PSERVER":
if not current_endpoint:
print("need env SERVER_ENDPOINT")
exit(1)
pserver_prog = t.get_pserver_program(current_endpoint)
pserver_startup = t.get_startup_program(current_endpoint,
pserver_prog)
print("starting server side startup")
exe.run(pserver_startup)
print("starting parameter server...")
exe.run(pserver_prog)
elif training_role == "TRAINER":
# Parameter initialization
exe.run(fluid.default_startup_program())
# data reader
train_reader = paddle.batch(
paddle.reader.shuffle(
paddle.dataset.cifar.train10() if args.data_set == 'cifar10'
else paddle.dataset.flowers.train(),
buf_size=5120),
batch_size=args.batch_size)
test_reader = paddle.batch(
paddle.dataset.cifar.test10() if args.data_set == 'cifar10' else
paddle.dataset.flowers.test(),
batch_size=args.batch_size)
trainer_prog = t.get_trainer_program()
feeder = fluid.DataFeeder(feed_list=[images, label], place=place)
# TODO(typhoonzero): change trainer startup program to fetch parameters from pserver
exe.run(fluid.default_startup_program())
train_loop(exe, trainer_prog)
else:
print("environment var TRAINER_ROLE should be TRAINER os PSERVER")
def print_arguments():
print('----------- Configuration Arguments -----------')
for arg, value in sorted(vars(args).iteritems()):
print('%s: %s' % (arg, value))
print('------------------------------------------------')
if __name__ == "__main__":
print_arguments()
main()
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#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 gzip
import paddle.v2.dataset.cifar as cifar
import paddle.v2 as paddle
import time
import os
DATA_DIM = 3 * 32 * 32
CLASS_DIM = 10
BATCH_SIZE = os.getenv("BATCH_SIZE")
if BATCH_SIZE:
BATCH_SIZE = int(BATCH_SIZE)
else:
BATCH_SIZE = 128
print "batch_size", BATCH_SIZE
NODE_COUNT = int(os.getenv("TRAINERS"))
ts = 0
def vgg(input, nums, class_dim):
def conv_block(input, num_filter, groups, num_channels=None):
return paddle.networks.img_conv_group(
input=input,
num_channels=num_channels,
pool_size=2,
pool_stride=2,
conv_num_filter=[num_filter] * groups,
conv_filter_size=3,
conv_act=paddle.activation.Relu(),
pool_type=paddle.pooling.Max())
assert len(nums) == 5
# the channel of input feature is 3
conv1 = conv_block(input, 64, nums[0], 3)
conv2 = conv_block(conv1, 128, nums[1])
conv3 = conv_block(conv2, 256, nums[2])
conv4 = conv_block(conv3, 512, nums[3])
conv5 = conv_block(conv4, 512, nums[4])
fc_dim = 512
fc1 = paddle.layer.fc(input=conv5,
size=fc_dim,
act=paddle.activation.Relu(),
layer_attr=paddle.attr.Extra(drop_rate=0.5))
fc2 = paddle.layer.fc(input=fc1,
size=fc_dim,
act=paddle.activation.Relu(),
layer_attr=paddle.attr.Extra(drop_rate=0.5))
out = paddle.layer.fc(input=fc2,
size=class_dim,
act=paddle.activation.Softmax())
return out
def vgg13(input, class_dim):
nums = [2, 2, 2, 2, 2]
return vgg(input, nums, class_dim)
def vgg16(input, class_dim):
nums = [2, 2, 3, 3, 3]
return vgg(input, nums, class_dim)
def vgg19(input, class_dim):
nums = [2, 2, 4, 4, 4]
return vgg(input, nums, class_dim)
def main():
global ts
paddle.init(use_gpu=False)
image = paddle.layer.data(
name="image", type=paddle.data_type.dense_vector(DATA_DIM))
lbl = paddle.layer.data(
name="label", type=paddle.data_type.integer_value(CLASS_DIM))
extra_layers = None
# NOTE: for v2 distributed training need averaging updates.
learning_rate = 1e-3 / NODE_COUNT
out = vgg16(image, class_dim=CLASS_DIM)
cost = paddle.layer.classification_cost(input=out, label=lbl)
# Create parameters
parameters = paddle.parameters.create(cost)
# Create optimizer
optimizer = paddle.optimizer.Momentum(
momentum=0.9,
regularization=paddle.optimizer.L2Regularization(rate=0.0005 *
BATCH_SIZE),
learning_rate=learning_rate / BATCH_SIZE,
learning_rate_decay_a=0.1,
learning_rate_decay_b=128000 * 35,
learning_rate_schedule="discexp", )
train_reader = paddle.batch(
paddle.reader.shuffle(
cifar.train10(),
# To use other data, replace the above line with:
# reader.train_reader('train.list'),
buf_size=1000),
batch_size=BATCH_SIZE)
test_reader = paddle.batch(
cifar.test10(),
# To use other data, replace the above line with:
# reader.test_reader('val.list'),
batch_size=BATCH_SIZE)
# Create trainer
trainer = paddle.trainer.SGD(cost=cost,
parameters=parameters,
update_equation=optimizer,
extra_layers=extra_layers,
is_local=False)
# End batch and end pass event handler
def event_handler(event):
global ts, ts_pass
if isinstance(event, paddle.event.BeginPass):
ts_pass = time.time()
if isinstance(event, paddle.event.BeginIteration):
ts = time.time()
if isinstance(event, paddle.event.EndIteration):
if event.batch_id % 1 == 0:
print "\nPass %d, Batch %d, Cost %f, %s, spent: %f" % (
event.pass_id, event.batch_id, event.cost, event.metrics,
time.time() - ts)
if isinstance(event, paddle.event.EndPass):
print "Pass %d end, spent: %f" % (event.pass_id,
time.time() - ts_pass)
result = trainer.test(reader=test_reader)
print "\nTest with Pass %d, %s" % (event.pass_id, result.metrics)
trainer.train(
reader=train_reader, num_passes=200, event_handler=event_handler)
if __name__ == '__main__':
main()
......@@ -21,6 +21,7 @@ set(BOOST_URL "http://sourceforge.net/projects/boost/files/boost/${BOO
set(BOOST_SOURCES_DIR ${THIRD_PARTY_PATH}/boost)
set(BOOST_DOWNLOAD_DIR "${BOOST_SOURCES_DIR}/src/${BOOST_PROJECT}")
set(BOOST_INCLUDE_DIR "${BOOST_DOWNLOAD_DIR}/${BOOST_TAR}" CACHE PATH "boost include directory." FORCE)
set_directory_properties(PROPERTIES CLEAN_NO_CUSTOM 1)
include_directories(${BOOST_INCLUDE_DIR})
......
......@@ -28,9 +28,3 @@ endif()
add_dependencies(eigen3 extern_eigen3)
LIST(APPEND external_project_dependencies eigen3)
IF(NOT WITH_C_API AND WITH_FLUID)
INSTALL(FILES ${EIGEN_INCLUDE_DIR}/Eigen/Core DESTINATION third_party/eigen3/Eigen)
INSTALL(DIRECTORY ${EIGEN_INCLUDE_DIR}/Eigen/src DESTINATION third_party/eigen3/Eigen)
INSTALL(DIRECTORY ${EIGEN_INCLUDE_DIR}/unsupported/Eigen DESTINATION third_party/eigen3/unsupported)
ENDIF()
......@@ -52,7 +52,7 @@ ADD_DEPENDENCIES(gflags extern_gflags)
LIST(APPEND external_project_dependencies gflags)
IF(WITH_C_API OR WITH_FLUID)
IF(WITH_C_API)
INSTALL(DIRECTORY ${GFLAGS_INCLUDE_DIR} DESTINATION third_party/gflags)
IF(ANDROID)
INSTALL(FILES ${GFLAGS_LIBRARIES} DESTINATION third_party/gflags/lib/${ANDROID_ABI})
......
......@@ -68,7 +68,7 @@ LINK_LIBRARIES(glog gflags)
LIST(APPEND external_project_dependencies glog)
IF(WITH_C_API OR WITH_FLUID)
IF(WITH_C_API)
INSTALL(DIRECTORY ${GLOG_INCLUDE_DIR} DESTINATION third_party/glog)
IF(ANDROID)
INSTALL(FILES ${GLOG_LIBRARIES} DESTINATION third_party/glog/lib/${ANDROID_ABI})
......
......@@ -250,7 +250,7 @@ IF(NOT PROTOBUF_FOUND)
SET(PROTOBUF_PROTOC_LIBRARY ${extern_protobuf_PROTOC_LIBRARY}
CACHE FILEPATH "protoc library." FORCE)
IF(WITH_C_API OR WITH_FLUID)
IF(WITH_C_API)
INSTALL(DIRECTORY ${PROTOBUF_INCLUDE_DIR} DESTINATION third_party/protobuf)
IF(ANDROID)
INSTALL(FILES ${PROTOBUF_LITE_LIBRARY} DESTINATION third_party/protobuf/lib/${ANDROID_ABI})
......
......@@ -52,6 +52,7 @@ ExternalProject_Add(
-DWITH_TORCH=OFF
-DCMAKE_DISABLE_FIND_PACKAGE_Torch=ON
-DBUILD_SHARED=ON
-DBUILD_TESTS=OFF
-DCMAKE_POSITION_INDEPENDENT_CODE=ON
-DCMAKE_BUILD_TYPE=${THIRD_PARTY_BUILD_TYPE}
${EXTERNAL_OPTIONAL_ARGS}
......
# make package for paddle fluid shared and static library
function(copy TARGET)
set(options "")
set(oneValueArgs "")
set(multiValueArgs SRCS DSTS DEPS)
cmake_parse_arguments(copy_lib "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
list(LENGTH copy_lib_SRCS copy_lib_SRCS_len)
list(LENGTH copy_lib_DSTS copy_lib_DSTS_len)
if(NOT ${copy_lib_SRCS_len} EQUAL ${copy_lib_DSTS_len})
message(FATAL_ERROR "${TARGET} source numbers are not equal to destination numbers")
endif()
math(EXPR len "${copy_lib_SRCS_len} - 1")
add_custom_target(${TARGET} DEPENDS ${copy_lib_DEPS})
foreach(index RANGE ${len})
list(GET copy_lib_SRCS ${index} src)
list(GET copy_lib_DSTS ${index} dst)
add_custom_command(TARGET ${TARGET} PRE_BUILD COMMAND mkdir -p "${dst}")
if(IS_DIRECTORY ${src})
add_custom_command(TARGET ${TARGET} PRE_BUILD COMMAND cp -r "${src}" "${dst}")
else()
add_custom_command(TARGET ${TARGET} PRE_BUILD COMMAND cp "${src}" "${dst}")
endif()
endforeach()
endfunction()
# third party
set(dst_dir "${CMAKE_INSTALL_PREFIX}/third_party/eigen3")
copy(eigen3_lib
SRCS ${EIGEN_INCLUDE_DIR}/Eigen/Core ${EIGEN_INCLUDE_DIR}/Eigen/src ${EIGEN_INCLUDE_DIR}/unsupported/Eigen
DSTS ${dst_dir}/Eigen ${dst_dir}/Eigen ${dst_dir}/unsupported
)
set(dst_dir "${CMAKE_INSTALL_PREFIX}/third_party/install/gflags")
copy(gflags_lib
SRCS ${GFLAGS_INCLUDE_DIR} ${GFLAGS_LIBRARIES}
DSTS ${dst_dir} ${dst_dir}/lib
)
set(dst_dir "${CMAKE_INSTALL_PREFIX}/third_party/install/glog")
copy(glog_lib
SRCS ${GLOG_INCLUDE_DIR} ${GLOG_LIBRARIES}
DSTS ${dst_dir} ${dst_dir}/lib
)
IF(NOT PROTOBUF_FOUND)
set(dst_dir "${CMAKE_INSTALL_PREFIX}/third_party/install/protobuf")
copy(protobuf_lib
SRCS ${PROTOBUF_INCLUDE_DIR} ${PROTOBUF_LITE_LIBRARY}
DSTS ${dst_dir} ${dst_dir}/lib
)
ENDIF(NOT PROTOBUF_FOUND)
# paddle fluid module
set(src_dir "${PADDLE_SOURCE_DIR}/paddle")
set(dst_dir "${CMAKE_INSTALL_PREFIX}/paddle")
set(module "framework")
copy(framework_lib DEPS framework_py_proto
SRCS ${src_dir}/${module}/*.h ${src_dir}/${module}/details/*.h ${PADDLE_BINARY_DIR}/paddle/framework/framework.pb.h
DSTS ${dst_dir}/${module} ${dst_dir}/${module}/details ${dst_dir}/${module}
)
set(module "memory")
copy(memory_lib
SRCS ${src_dir}/${module}/*.h ${src_dir}/${module}/detail/*.h
DSTS ${dst_dir}/${module} ${dst_dir}/${module}/detail
)
set(module "inference")
copy(inference_lib DEPENDS paddle_fluid_shared
SRCS ${src_dir}/${module}/*.h ${PADDLE_BINARY_DIR}/paddle/inference/libpaddle_fluid.so
DSTS ${dst_dir}/${module} ${dst_dir}/${module}
)
set(module "platform")
copy(platform_lib
SRCS ${src_dir}/${module}/*.h ${src_dir}/${module}/dynload/*.h ${src_dir}/${module}/details/*.h
DSTS ${dst_dir}/${module} ${dst_dir}/${module}/dynload ${dst_dir}/${module}/details
)
set(module "string")
copy(string_lib
SRCS ${src_dir}/${module}/*.h ${src_dir}/${module}/tinyformat/*.h
DSTS ${dst_dir}/${module} ${dst_dir}/${module}/tinyformat
)
add_custom_target(inference_lib_dist DEPENDS
inference_lib framework_lib memory_lib platform_lib string_lib
gflags_lib glog_lib protobuf_lib eigen3_lib)
......@@ -47,3 +47,5 @@ sphinx_add_target(paddle_docs_cn
${SPHINX_CACHE_DIR_CN}
${CMAKE_CURRENT_SOURCE_DIR}
${SPHINX_HTML_DIR_CN})
add_subdirectory(api)
# configured documentation tools and intermediate build results
set(BINARY_BUILD_DIR_EN "${CMAKE_CURRENT_BINARY_DIR}/en/_build")
# Sphinx cache with pickled ReST documents
set(SPHINX_CACHE_DIR_EN "${CMAKE_CURRENT_BINARY_DIR}/en/_doctrees")
# HTML output director
set(SPHINX_HTML_DIR_EN "${CMAKE_CURRENT_BINARY_DIR}/en/html")
configure_file(
"${CMAKE_CURRENT_SOURCE_DIR}/../templates/conf.py.en.in"
"${BINARY_BUILD_DIR_EN}/conf.py"
@ONLY)
sphinx_add_target(paddle_api_docs
html
${BINARY_BUILD_DIR_EN}
${SPHINX_CACHE_DIR_EN}
${CMAKE_CURRENT_SOURCE_DIR}
${SPHINX_HTML_DIR_EN})
......@@ -87,6 +87,11 @@ roi_pool
.. autoclass:: paddle.v2.layer.roi_pool
:noindex:
pad
----
.. autoclass:: paddle.v2.layer.pad
:noindex:
Norm Layer
==========
......@@ -133,6 +138,11 @@ grumemory
.. autoclass:: paddle.v2.layer.grumemory
:noindex:
gated_unit
-----------
.. autoclass:: paddle.v2.layer.gated_unit
:noindex:
Recurrent Layer Group
=====================
......@@ -340,6 +350,11 @@ bilinear_interp
.. autoclass:: paddle.v2.layer.bilinear_interp
:noindex:
dropout
--------
.. autoclass:: paddle.v2.layer.dropout
:noindex:
dot_prod
---------
.. autoclass:: paddle.v2.layer.dot_prod
......@@ -402,6 +417,11 @@ scale_shift
.. autoclass:: paddle.v2.layer.scale_shift
:noindex:
factorization_machine
---------------------
.. autoclass:: paddle.v2.layer.factorization_machine
:noindex:
Sampling Layers
===============
......@@ -420,22 +440,6 @@ multiplex
.. autoclass:: paddle.v2.layer.multiplex
:noindex:
Factorization Machine Layer
============================
factorization_machine
---------------------
.. autoclass:: paddle.v2.layer.factorization_machine
:noindex:
Slicing and Joining Layers
==========================
pad
----
.. autoclass:: paddle.v2.layer.pad
:noindex:
.. _api_v2.layer_costs:
Cost Layers
......@@ -526,6 +530,11 @@ multibox_loss
.. autoclass:: paddle.v2.layer.multibox_loss
:noindex:
detection_output
----------------
.. autoclass:: paddle.v2.layer.detection_output
:noindex:
Check Layer
============
......@@ -534,31 +543,10 @@ eos
.. autoclass:: paddle.v2.layer.eos
:noindex:
Miscs
=====
dropout
--------
.. autoclass:: paddle.v2.layer.dropout
:noindex:
Activation with learnable parameter
===================================
Activation
==========
prelu
--------
.. autoclass:: paddle.v2.layer.prelu
:noindex:
gated_unit
-----------
.. autoclass:: paddle.v2.layer.gated_unit
:noindex:
Detection output Layer
======================
detection_output
----------------
.. autoclass:: paddle.v2.layer.detection_output
:noindex:
......@@ -73,3 +73,10 @@ wmt14
.. automodule:: paddle.v2.dataset.wmt14
:members:
:noindex:
wmt16
+++++
.. automodule:: paddle.v2.dataset.wmt16
:members:
:noindex:
......@@ -13,7 +13,7 @@ PaddlePaddle提供pip和Docker的安装方式:
pip_install_cn.rst
docker_install_cn.rst
../../howto/dev/build_cn.md
build_cn.md
编译流程
++++++++
......
......@@ -13,7 +13,7 @@ You can choose either pip or Docker to complete your install:
pip_install_en.rst
docker_install_en.rst
../../howto/dev/build_en.md
build_en.md
Build from Source
......
## Auto Gradient Checker Design
## Auto Gradient Check Design
## Backgraound:
- Generally, it is easy to check whether the forward computation of an Operator is correct or not. However, backpropagation is a notoriously difficult algorithm to debug and get right:
1. you should get the right backpropagation formula according to the forward computation.
2. you should implement it right in CPP.
3. it's difficult to prepare test data.
## Background:
- Generally, it is easy to check whether the forward computation of an Operator is correct or not. However, backpropagation is a notoriously difficult algorithm to debug and get right because of the following challenges:
1. The formula for backpropagation formula should be correct according to the forward computation.
2. The Implementation of the above shoule be correct in CPP.
3. It is difficult to prepare an unbiased test data.
- Auto gradient checking gets a numerical gradient by forward Operator and use it as a reference of the backward Operator's result. It has several advantages:
1. numerical gradient checker only need forward operator.
2. user only need to prepare the input data for forward Operator.
- Auto gradient checking gets a numerical gradient using forward Operator and uses it as a reference for the backward Operator's result. It has several advantages:
1. Numerical gradient checker only needs the forward operator.
2. The user only needs to prepare the input data for forward Operator and not worry about the backward Operator.
## Mathematical Theory
The following two document from Stanford has a detailed explanation of how to get numerical gradient and why it's useful.
The following documents from Stanford have a detailed explanation of how to compute the numerical gradient and why it is useful.
- [Gradient checking and advanced optimization(en)](http://deeplearning.stanford.edu/wiki/index.php/Gradient_checking_and_advanced_optimization)
- [Gradient checking and advanced optimization(cn)](http://ufldl.stanford.edu/wiki/index.php/%E6%A2%AF%E5%BA%A6%E6%A3%80%E9%AA%8C%E4%B8%8E%E9%AB%98%E7%BA%A7%E4%BC%98%E5%8C%96)
## Numeric Gradient Implementation
## Numerical Gradient Implementation
### Python Interface
```python
def get_numerical_gradient(op,
......@@ -27,73 +27,76 @@ def get_numerical_gradient(op,
delta=0.005,
local_scope=None):
"""
Get Numeric Gradient for an operator's input.
Get Numerical Gradient for the input of an operator.
:param op: C++ operator instance, could be an network
:param op: C++ operator instance, could be an network.
:param input_values: The input variables. Should be an dictionary, whose key is
variable name, and value is numpy array.
variable name, and value is a numpy array.
:param output_name: The final output variable name.
:param input_to_check: The input variable with respect to which to compute the gradient.
:param delta: The perturbation value for numeric gradient method. The
smaller delta is, the more accurate result will get. But if that delta is
too small, it will suffer from numerical stability problem.
:param input_to_check: The input variable with respect to which the gradient has to be computed.
:param delta: The perturbation value for numerical gradient method. The
smaller the delta, the more accurate the result. But if the delta is too
small, it will suffer from the numerical stability problem.
:param local_scope: The local scope used for get_numeric_gradient.
:return: The gradient array in numpy format.
"""
```
### Explaination:
### Explanation:
- Why need `output_name`
- An Operator may have multiple Output, one can get independent gradient from each Output. So caller should specify the name of the output variable.
- Why do we need an `output_name`
- An Operator may have multiple Outputs, one can compute an independent gradient from each Output. So the caller should specify the name of the output variable.
- Why need `input_to_check`
- One operator may have multiple inputs. Gradient Op can calculate the gradient of these inputs at the same time. But Numeric Gradient needs to calculate them one by one. So `get_numeric_gradient` is designed to calculate the gradient for one input. If you need to compute multiple inputs, you can call `get_numeric_gradient` multiple times.
- Why do we need `input_to_check`
- One operator can have multiple inputs. Gradient Op can calculate the gradient of these inputs at the same time. But Numerical Gradient needs to calculate them one by one. So `get_numeric_gradient` is designed to calculate the gradient for one input. If you need to compute multiple inputs, you can call `get_numeric_gradient` multiple times each with a different input.
### Core Algorithm Implementation
```python
# we only compute gradient of one element a time.
# we only compute the gradient of one element a time.
# we use a for loop to compute the gradient of each element.
for i in xrange(tensor_size):
# get one input element by its index i.
origin = tensor_to_check.get_float_element(i)
# get one input element using the index i.
original = tensor_to_check.get_float_element(i)
# add delta to it, run op and then get the new value of the result tensor.
x_pos = origin + delta
# add delta to it, run the forward op and then
# get the new value of the result tensor.
x_pos = original + delta
tensor_to_check.set_float_element(i, x_pos)
y_pos = get_output()
# plus delta to this element, run op and get the new value of the result tensor.
x_neg = origin - delta
# Subtract delta from this element, run the op again
# and get the new value of the result tensor.
x_neg = original - delta
tensor_to_check.set_float_element(i, x_neg)
y_neg = get_output()
# restore old value
tensor_to_check.set_float_element(i, origin)
tensor_to_check.set_float_element(i, original)
# compute the gradient of this element and store it into a numpy array.
# compute the gradient of this element and store
# it into a numpy array.
gradient_flat[i] = (y_pos - y_neg) / delta / 2
# reshape the gradient result to the shape of the source tensor.
return gradient_flat.reshape(tensor_to_check.get_dims())
```
## Auto Graident Checker Framework
## Auto Gradient Check Framework
Each Operator Kernel has three kinds of Gradient:
1. Numerical gradient
2. CPU kernel gradient
3. GPU kernel gradient (if supported)
3. GPU kernel gradient (if supported by the device)
The numerical gradient only relies on forward Operator. So we use the numerical gradient as the reference value. And the gradient checking is performed in the following three steps:
The numerical gradient only relies on the forward Operator, so we use the numerical gradient as the reference value. The gradient checking is performed in the following three steps:
1. calculate the numerical gradient
2. calculate CPU kernel gradient with the backward Operator and compare it with the numerical gradient
3. calculate GPU kernel gradient with the backward Operator and compare it with the numeric gradient (if supported)
1. Calculate the numerical gradient
2. Calculate CPU kernel gradient with the backward Operator and compare it with the numerical gradient.
3. Calculate GPU kernel gradient with the backward Operator and compare it with the numeric gradient. (if supported)
#### Python Interface
......@@ -109,26 +112,27 @@ The numerical gradient only relies on forward Operator. So we use the numerical
"""
:param forward_op: used to create backward_op
:param input_vars: numpy value of input variable. The following
computation will use these variables.
:param inputs_to_check: the input variable with respect to which to compute the gradient.
computation will use these variables.
:param inputs_to_check: the input variable with respect to which the
gradient will be computed.
:param output_name: The final output variable name.
:param max_relative_error: The relative tolerance parameter.
:param no_grad_set: used when create backward ops
:param no_grad_set: used to create backward ops
:param only_cpu: only compute and check gradient on cpu kernel.
:return:
"""
```
### How to check if two numpy array is close enough?
if `abs_numerical_grad` is nearly zero, then use abs error for numerical_grad
### How to check if two numpy arrays are close enough?
if `abs_numerical_grad` is nearly zero, then use absolute error for numerical_grad.
```python
numerical_grad = ...
operator_grad = numpy.array(scope.find_var(grad_var_name(name)).get_tensor())
abs_numerical_grad = numpy.abs(numerical_grad)
# if abs_numerical_grad is nearly zero, then use abs error for numeric_grad, not relative
# error.
# if abs_numerical_grad is nearly zero, then use abs error for
# numeric_grad, instead of relative error.
abs_numerical_grad[abs_numerical_grad < 1e-3] = 1
diff_mat = numpy.abs(abs_numerical_grad - operator_grad) / abs_numerical_grad
......@@ -137,10 +141,10 @@ max_diff = numpy.max(diff_mat)
#### Notes:
The Input data for auto gradient checker should be reasonable to avoid numerical stability problem.
The Input data for auto gradient checker should be reasonable to avoid numerical stability problem.
#### Refs:
#### References:
- [Gradient checking and advanced optimization(en)](http://deeplearning.stanford.edu/wiki/index.php/Gradient_checking_and_advanced_optimization)
- [Gradient checking and advanced optimization(cn)](http://ufldl.stanford.edu/wiki/index.php/%E6%A2%AF%E5%BA%A6%E6%A3%80%E9%AA%8C%E4%B8%8E%E9%AB%98%E7%BA%A7%E4%BC%98%E5%8C%96)
......@@ -42,7 +42,7 @@ The type *channel* is conceptually the blocking queue. In Go, its implemented i
The `select` operation has been in OS kernels long before Go language. All Unix kernels implement system calls *poll* and *select*. They monitor multiple file descriptors to see if I/O is possible on any of them. This takes O(N) time. Since Linux 2.6, a new system call, *epoll*, can do the same in O(1) time. In BSD systems, there is a similar system call *kqueue*. Go's Linux implementation uses epoll.
It might be a good idea to implement Fluid's select using epoll too. In this design doc, we start from the O(N) way, so we could focus on Python binding and the syntax.
It might be a good idea to implement Fluid's select using epoll too. In this design doc, we start from the O(N) way so that we could focus on Python binding and the syntax.
### Type Channel
......@@ -71,14 +71,14 @@ ch1 := make(chan int, 100) // a channel that can buffer 100 ints.
In Fluid, we should be able to do the same:
```python
ch = fluid.make_chan(dtype=INT)
ch1 = fluid.make_chan(dtype=INT, 100)
ch = fluid.make_channel(dtype=INT)
ch1 = fluid.make_channel(dtype=INT, 100)
```
In addition to that, we want channels that can hold more complex element types, e.g., Tensors of float16:
```python
ch = fluid.make_chan(dtype=Tensor, etype=float16)
ch = fluid.make_channel(dtype=Tensor, etype=float16)
```
or Tensors of Tensors of float16 etc.
......@@ -87,8 +87,136 @@ The point here is that we need a consistent way to compose types, like in C++ we
### Send and Recv
Go's CSP implementation depends on data type *channel*. There are two types of channels:
1. The unblocked channel, or buffered channel, is a blocking queue with a non-zero sized buffer. The sending to buffered channel blocks if the buffer is full, and the receive operation blocks if the buffer is empty.
1. blocked channel, or unbuffered channel, is a blocking queue with no buffer. Both sending and receiving block with unbuffered channels.
There are four types of actions with a channel:
1. Create a channel
```go
ch := make(chan int) // this is an unbuffered channel
ch := make(chan int, 100) // this is a buffered channel of 100 ints.
```
1. Send
```go
ch <- 111
```
1. Recv
```go
y, ok <- ch
```
1. Close
```go
close(ch)
```
Please be aware that a closed channel is not a nil channel, which is `var ch chan int`.
There are some [axioms with channels](https://dave.cheney.net/2014/03/19/channel-axioms):
1. A send to a nil channel blocks forever
1. A receive from a nil channel blocks forever
1. A send to a closed channel panics
1. A receive from a closed channel returns the residual values and then zeros.
In Fluid, we have [buffered channels](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/details/buffered_channel.h) and [unbuffered channels](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/details/unbuffered_channel.h)
The following program illustrates the Python syntax for accessing Fluid buffers.
```python
import fluid
buffer_size = 10
ch = fluid.make_channel(dtype=INT, buffer_size)
# Now write three elements to the channel
with fluid.while(steps=buffer_size):
fluid.send(ch, step)
fluid.close_channel(ch)
with fluid.while(steps=buffer_size):
fluid.print(fluid.recv(ch))
```
The following example shows that to avoid the always-blocking behavior of unbuffered channels, we need to use Fluid's goroutines.
```python
import fluid
ch = fluid.make_channel(dtype=INT)
with fluid.go():
fluid.send(ch)
y = fluid.recv(ch)
fluid.close_channel(ch)
```
### Select
In Go, the `select` statement lets a goroutine wait on multiple communication operations. A `select` blocks until one of its cases can run, then it executes that case. It chooses one at random if multiple are ready.
```go
ch1 := make(chan int)
ch2 := make(chan int, 100)
x := 0
for {
select {
case ch1 <- x:
x := x + 1
case y <- ch2:
fmt.Println("Received on channel")
default:
fmt.Println("Default")
}
}
```
In Fluid, we should be able to do the same:
```python
ch1 = fluid.make_chan(dtype=INT)
ch2 = fluid.make_chan(dtype=INT, 100)
sel = fluid.select()
with sel.case(ch1, 'w', X):
fluid.layers.increment(X)
with sel.case(ch2, 'r', Y):
fluid.print("Received on Channel")
with sel.default():
fluid.print("Default")
```
In the above code snippet, `X` and `Y` are variables. Now let us look at each of these statements one by one.
- `sel.case(ch1, 'w', X)` : This specifies that we are writing to `ch1` and we want to write the integer in variable `X` to the channel. The character `w` is used here to make the syntax familiar to write syntax in Python I/O.
- `sel.case(ch2, 'r', Y)` : This specifies that we would like to read the result from `ch2` into variable `Y`. The character `r` is used here to make the syntax familiar to read syntax in Python I/O.
- `sel.default()` : This is equivalent to the default in Go `select`. If none of the channels are ready for read or write, then the fluid code in the default block will be executed.
## Example Programs
### 1. RPC between Trainers and Parameter Servers
......
### Design Doc: Switch
### Background
Many programming languages provide `switch` as a generalization of `if-elif-else`. We want to add it to Fluid.
The following example shows the usage of `fluid.switch`.
```python
a = fluid.Var(10)
b = fluid.Var(0)
with switch() as switch:
with switch.case(fluid.less_equal(a, 10)):
fluid.print("Case 1")
with switch.case(fluid.larger(a, 0)):
fluid.print("Case 2")
with switch.default():
fluid.print("Case 3")
```
### The Semantics
1. A `switch` control-flow checks cases one-by-one.
1. The condition of each case is a boolean value, which is a scalar, and differs from the `fluid.if_else` control-flow, which condition could be a vector of boolean values.
1. It runs the first matched case, or the default case if there is one.
1. Once it matches a case, it runs the corresponding branch and only that branch. It's like there is a C's `break` keyword at the end of each case.
The above program should print and print only "Case 1".
The implementation of the backward pass of the `switch` control-flow is easier than the backward of the `if_else`, because `switch` runs at most one branch, whereas `if-else` could run more than one branches.
../../CONTRIBUTING.md
\ No newline at end of file
开发标准
========
.. toctree::
:maxdepth: 1
contribute_to_paddle_cn.md
write_docs_cn.rst
Development
------------
.. toctree::
:maxdepth: 1
new_layer_en.rst
contribute_to_paddle_en.md
write_docs_en.rst
##################
如何贡献/修改文档
##################
#############
如何贡献文档
#############
PaddlePaddle的文档包括英文文档 ``doc`` 和中文文档 ``doc_cn`` 两个部分。文档都是通过 `cmake`_ 驱动 `sphinx`_ 编译生成,生成后的文档分别存储在编译目录的 ``doc`` 和 ``doc_cn`` 两个子目录下。
也可以利用PaddlePaddle 工具来编译文档,这个情况下所有的文件会存在整理过的的文件目录 .ppo_workspace/content 下
......
##################
########################
Contribute Documentation
##################
########################
PaddlePaddle supports English documentation ``doc`` and Chinese documentation ``doc_cn``.
Both are compiled by `cmake`_ and `sphinx`_ , the compiled documentations will be stored under ``doc`` and ``doc_cn`` directories.
......
......@@ -4,7 +4,7 @@
PaddlePaddle是源于百度的一个深度学习平台。PaddlePaddle为深度学习研究人员提供了丰富的API,可以轻松地完成神经网络配置,模型训练等任务。
这里将介绍PaddlePaddle的基本使用概念,并且展示了如何利用PaddlePaddle来解决一个经典的线性回归问题。
在使用该文档之前,请参考 `安装文档 <../build_and_install/index_cn.html>`_ 完成PaddlePaddle的安装。
在使用该文档之前,请参考 `安装文档 <../../build_and_install/index_cn.html>`_ 完成PaddlePaddle的安装。
配置网络
......
新手入门
============
.. _quick_install:
快速安装
++++++++
PaddlePaddle支持使用pip快速安装,目前支持CentOS 6以上, Ubuntu 14.04以及MacOS 10.12,并安装有Python2.7。
执行下面的命令完成快速安装,版本为cpu_avx_openblas:
.. code-block:: bash
pip install paddlepaddle
如果需要安装支持GPU的版本(cuda7.5_cudnn5_avx_openblas),需要执行:
.. code-block:: bash
pip install paddlepaddle-gpu
更详细的安装和编译方法参考:
.. toctree::
:maxdepth: 1
build_and_install/index_cn.rst
.. _quick_start:
快速开始
++++++++
创建一个 housing.py 并粘贴此Python代码:
.. code-block:: python
import paddle.v2 as paddle
# Initialize PaddlePaddle.
paddle.init(use_gpu=False, trainer_count=1)
# Configure the neural network.
x = paddle.layer.data(name='x', type=paddle.data_type.dense_vector(13))
y_predict = paddle.layer.fc(input=x, size=1, act=paddle.activation.Linear())
# Infer using provided test data.
probs = paddle.infer(
output_layer=y_predict,
parameters=paddle.dataset.uci_housing.model(),
input=[item for item in paddle.dataset.uci_housing.test()()])
for i in xrange(len(probs)):
print 'Predicted price: ${:,.2f}'.format(probs[i][0] * 1000)
执行 :code:`python housing.py` 瞧! 它应该打印出预测住房数据的清单。
.. toctree::
:maxdepth: 1
quickstart_cn.rst
concepts/use_concepts_cn.rst
GET STARTED
============
.. _quick_install:
Quick Install
----------------------
You can use pip to install PaddlePaddle with a single command, supports
CentOS 6 above, Ubuntu 14.04 above or MacOS 10.12, with Python 2.7 installed.
Simply run the following command to install, the version is cpu_avx_openblas:
.. code-block:: bash
pip install paddlepaddle
If you need to install GPU version (cuda7.5_cudnn5_avx_openblas), run:
.. code-block:: bash
pip install paddlepaddle-gpu
For more details about installation and build:
.. toctree::
:maxdepth: 1
build_and_install/index_en.rst
.. _quick_start:
Quick Start
++++++++
Create a new file called housing.py, and paste this Python
code:
.. code-block:: python
import paddle.v2 as paddle
# Initialize PaddlePaddle.
paddle.init(use_gpu=False, trainer_count=1)
# Configure the neural network.
x = paddle.layer.data(name='x', type=paddle.data_type.dense_vector(13))
y_predict = paddle.layer.fc(input=x, size=1, act=paddle.activation.Linear())
# Infer using provided test data.
probs = paddle.infer(
output_layer=y_predict,
parameters=paddle.dataset.uci_housing.model(),
input=[item for item in paddle.dataset.uci_housing.test()()])
for i in xrange(len(probs)):
print 'Predicted price: ${:,.2f}'.format(probs[i][0] * 1000)
Run :code:`python housing.py` and voila! It should print out a list of predictions
for the test housing data.
quickstart_en.rst
快速开始
========
PaddlePaddle支持使用pip快速安装,目前支持CentOS 6以上, Ubuntu 14.04以及MacOS 10.12,并安装有Python2.7。
执行下面的命令完成快速安装,版本为cpu_avx_openblas:
.. code-block:: bash
pip install paddlepaddle
如果需要安装支持GPU的版本(cuda7.5_cudnn5_avx_openblas),需要执行:
.. code-block:: bash
pip install paddlepaddle-gpu
更详细的安装和编译方法参考::ref:`install_steps` 。
创建一个 housing.py 并粘贴此Python代码:
.. code-block:: python
import paddle.v2 as paddle
# Initialize PaddlePaddle.
paddle.init(use_gpu=False, trainer_count=1)
# Configure the neural network.
x = paddle.layer.data(name='x', type=paddle.data_type.dense_vector(13))
y_predict = paddle.layer.fc(input=x, size=1, act=paddle.activation.Linear())
# Infer using provided test data.
probs = paddle.infer(
output_layer=y_predict,
parameters=paddle.dataset.uci_housing.model(),
input=[item for item in paddle.dataset.uci_housing.test()()])
for i in xrange(len(probs)):
print 'Predicted price: ${:,.2f}'.format(probs[i][0] * 1000)
执行 :code:`python housing.py` 瞧! 它应该打印出预测住房数据的清单。
Quick Start
============
You can use pip to install PaddlePaddle with a single command, supports
CentOS 6 above, Ubuntu 14.04 above or MacOS 10.12, with Python 2.7 installed.
Simply run the following command to install, the version is cpu_avx_openblas:
.. code-block:: bash
pip install paddlepaddle
If you need to install GPU version (cuda7.5_cudnn5_avx_openblas), run:
.. code-block:: bash
pip install paddlepaddle-gpu
For more details about installation and build: :ref:`install_steps` .
Create a new file called housing.py, and paste this Python
code:
.. code-block:: python
import paddle.v2 as paddle
# Initialize PaddlePaddle.
paddle.init(use_gpu=False, trainer_count=1)
# Configure the neural network.
x = paddle.layer.data(name='x', type=paddle.data_type.dense_vector(13))
y_predict = paddle.layer.fc(input=x, size=1, act=paddle.activation.Linear())
# Infer using provided test data.
probs = paddle.infer(
output_layer=y_predict,
parameters=paddle.dataset.uci_housing.model(),
input=[item for item in paddle.dataset.uci_housing.test()()])
for i in xrange(len(probs)):
print 'Predicted price: ${:,.2f}'.format(probs[i][0] * 1000)
Run :code:`python housing.py` and voila! It should print out a list of predictions
for the test housing data.
## 编译 PaddlePaddle 预测库
## 安装与编译C-API预测库
### 概述
......
PaddlePaddle C-API
C-API预测库
==================
.. toctree::
......
## C-API 使用流程
## C-API使用流程
这篇文档介绍 PaddlePaddle C-API 整体使用流程。
......
# 分布式训练
## 概述
本文将介绍如何使用PaddlePaddle在不同的集群框架下完成分布式训练。分布式训练架构如下图所示:
<img src="https://user-images.githubusercontent.com/13348433/31772175-5f419eca-b511-11e7-9db7-5231fe3d9ccb.png" width="500">
- 数据分片(Data shard): 用于训练神经网络的数据,被切分成多个部分,每个部分分别给每个trainer使用。
- 计算节点(Trainer): 每个trainer启动后读取切分好的一部分数据,开始神经网络的“前馈”和“后馈”计算,并和参数服务器通信。在完成一定量数据的训练后,上传计算得出的梯度(gradients),然后下载优化更新后的神经网络参数(parameters)。
- 参数服务器(Parameter server):每个参数服务器只保存整个神经网络所有参数的一部分。参数服务器接收从计算节点上传的梯度,并完成参数优化更新,再将更新后的参数下发到每个计算节点。
这样,通过计算节点和参数服务器的分布式协作,可以完成神经网络的SGD方法的训练。PaddlePaddle可以同时支持同步随机梯度下降(SGD)和异步随机梯度下降。
在使用同步SGD训练神经网络时,PaddlePaddle使用同步屏障(barrier),使梯度的提交和参数的更新按照顺序方式执行。在异步SGD中,则并不会等待所有trainer提交梯度才更新参数,这样极大地提高了计算的并行性:参数服务器之间不相互依赖,并行地接收梯度和更新参数,参数服务器也不会等待计算节点全部都提交梯度之后才开始下一步,计算节点之间也不会相互依赖,并行地执行模型的训练。可以看出,虽然异步SGD方式会提高参数更新并行度, 但是并不能保证参数同步更新,在任意时间某一台参数服务器上保存的参数可能比另一台要更新,与同步SGD相比,梯度会有噪声。
## 环境准备
1. 准备您的计算集群。计算集群通常由一组(几台到几千台规模)的Linux服务器组成。服务器之间可以通过局域网(LAN)联通,每台服务器具有集群中唯一的IP地址(或者可被DNS解析的主机名)。集群中的每台计算机通常被成为一个“节点”。
1. 我们需要在集群的所有节点上安装 PaddlePaddle。 如果要启用GPU,还需要在节点上安装对应的GPU驱动以及CUDA。PaddlePaddle的安装可以参考[build_and_install](http://www.paddlepaddle.org/docs/develop/documentation/zh/getstarted/build_and_install/index_cn.html)的多种安装方式。我们推荐使用[Docker](http://www.paddlepaddle.org/docs/develop/documentation/zh/getstarted/build_and_install/docker_install_cn.html)安装方式来快速安装PaddlePaddle。
安装完成之后,执行下面的命令可以查看已经安装的版本(docker安装方式可以进入docker容器执行:`docker run -it paddlepaddle/paddle:[tag] /bin/bash`):
```bash
$ paddle version
PaddlePaddle 0.10.0, compiled with
with_avx: ON
with_gpu: OFF
with_double: OFF
with_python: ON
with_rdma: OFF
with_timer: OFF
```
## 启动参数说明
下面以`doc/howto/usage/cluster/src/word2vec`中的代码作为实例,介绍使用PaddlePaddle v2 API完成分布式训练。
下面以`doc/howto/cluster/src/word2vec`中的代码作为实例,介绍使用PaddlePaddle v2 API完成分布式训练。
## 启动参数说明
### 启动参数服务器
执行以下的命令启动一个参数服务器并等待和计算节点的数据交互
```bash
......@@ -167,22 +133,3 @@ test.txt-00002
- `train_data_dir`:包含训练数据的目录,可以是从分布式存储挂载过来的,也可以是在任务启动前下载到本地的。
- `test_data_dir`:包含测试数据集的目录。
## 使用分布式计算平台或工具
PaddlePaddle可以使用多种分布式计算平台构建分布式计算任务,包括:
- [Kubernetes](http://kubernetes.io) Google开源的容器集群的调度框架,支持大规模集群生产环境的完整集群方案。
- [OpenMPI](https://www.open-mpi.org) 成熟的高性能并行计算框架。
- [Fabric](http://www.fabfile.org) 集群管理工具。可以使用`Fabric`编写集群任务提交和管理脚本。
对于不同的集群平台,会分别介绍集群作业的启动和停止方法。这些例子都可以在[cluster_train_v2](https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/scripts/cluster_train_v2)找到。
在使用分布式计算平台进行训练时,任务被调度在集群中时,分布式计算平台通常会通过API或者环境变量提供任务运行需要的参数,比如节点的ID、IP和任务节点个数等。
## 在不同集群中运行
- [fabric集群](fabric_cn.md)
- [openmpi集群](openmpi_cn.md)
- [kubernetes单机](k8s_cn.md)
- [kubernetes distributed分布式](k8s_distributed_cn.md)
- [AWS上运行kubernetes集群训练](k8s_aws_cn.md)
# Distributed Training
## Introduction
In this article, we'll explain how to run distributed training jobs with PaddlePaddle on different types of clusters. The diagram below shows the main architecture of a distributed trainning job:
<img src="https://user-images.githubusercontent.com/13348433/31772146-41523d84-b511-11e7-8a12-a69fd136c283.png" width="500">
- Data shard: training data will be split into multiple partitions, trainers use the partitions of the whole dataset to do the training job.
- Trainer: each trainer reads the data shard, and train the neural network. Then the trainer will upload calculated "gradients" to parameter servers, and wait for parameters to be optimized on the parameter server side. When that finishes, the trainer download optimized parameters and continues its training.
- Parameter server: every parameter server stores part of the whole neural network model data. They will do optimization calculations when gradients are uploaded from trainers, and then send updated parameters to trainers.
PaddlePaddle can support both synchronize stochastic gradient descent (SGD) and asynchronous SGD.
When training with synchronize SGD, PaddlePaddle uses an internal "synchronize barrier" which makes gradients update and parameter download in strict order. On the other hand, asynchronous SGD won't wait for all trainers to finish upload at a single step, this will increase the parallelism of distributed training: parameter servers do not depend on each other, they'll do parameter optimization concurrently. Parameter servers will not wait for trainers, so trainers will also do their work concurrently. But asynchronous SGD will introduce more randomness and noises in the gradient.
## Preparations
1. Prepare your computer cluster. It's normally a bunch of Linux servers connected by LAN. Each server will be assigned a unique IP address. The computers in the cluster can be called "nodes".
2. Install PaddlePaddle on every node. If you are going to take advantage of GPU cards, you'll also need to install proper driver and CUDA libraries. To install PaddlePaddle please read [this build and install](http://www.paddlepaddle.org/docs/develop/documentation/en/getstarted/build_and_install/index_en.html) document. We strongly recommend using [Docker installation](http://www.paddlepaddle.org/docs/develop/documentation/en/getstarted/build_and_install/docker_install_en.html).
After installation, you can check the version by typing the below command (run a docker container if using docker: `docker run -it paddlepaddle/paddle:[tag] /bin/bash`):
```bash
$ paddle version
PaddlePaddle 0.10.0rc, compiled with
with_avx: ON
with_gpu: OFF
with_double: OFF
with_python: ON
with_rdma: OFF
with_timer: OFF
```
We'll take `doc/howto/usage/cluster/src/word2vec` as an example to introduce distributed training using PaddlePaddle v2 API.
## Command-line arguments
We'll take `doc/howto/cluster/src/word2vec` as an example to introduce distributed training using PaddlePaddle v2 API.
### Starting parameter server
Type the below command to start a parameter server which will wait for trainers to connect:
......@@ -171,21 +138,3 @@ Your workspace may looks like:
- `train_data_dir`: containing training data. Mount from storage service or copy trainning data to here.
- `test_data_dir`: containing testing data.
## Use cluster platforms or cluster management tools
PaddlePaddle supports running jobs on several platforms including:
- [Kubernetes](http://kubernetes.io) open-source system for automating deployment, scaling, and management of containerized applications from Google.
- [OpenMPI](https://www.open-mpi.org) Mature high performance parallel computing framework.
- [Fabric](http://www.fabfile.org) A cluster management tool. Write scripts to submit jobs or manage the cluster.
We'll introduce cluster job management on these platforms. The examples can be found under [cluster_train_v2](https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/scripts/cluster_train_v2).
These cluster platforms provide API or environment variables for training processes, when the job is dispatched to different nodes. Like node ID, IP or total number of nodes etc.
## Use different clusters
- [fabric](fabric_en.md)
- [openmpi](openmpi_en.md)
- [kubernetes](k8s_en.md)
- [kubernetes on AWS](k8s_aws_en.md)
分布式训练
==========
.. toctree::
:maxdepth: 1
introduction_cn.md
preparations_cn.md
cmd_argument_cn.md
multi_cluster/index_cn.rst
Distributed Training
====================
.. toctree::
:maxdepth: 1
introduction_en.md
preparations_en.md
cmd_argument_en.md
multi_cluster/index_en.rst
## 概述
本节将介绍如何使用PaddlePaddle在不同的集群框架下完成分布式训练。分布式训练架构如下图所示:
<img src="https://user-images.githubusercontent.com/13348433/31772175-5f419eca-b511-11e7-9db7-5231fe3d9ccb.png" width="500">
- 数据分片(Data shard): 用于训练神经网络的数据,被切分成多个部分,每个部分分别给每个trainer使用。
- 计算节点(Trainer): 每个trainer启动后读取切分好的一部分数据,开始神经网络的“前馈”和“后馈”计算,并和参数服务器通信。在完成一定量数据的训练后,上传计算得出的梯度(gradients),然后下载优化更新后的神经网络参数(parameters)。
- 参数服务器(Parameter server):每个参数服务器只保存整个神经网络所有参数的一部分。参数服务器接收从计算节点上传的梯度,并完成参数优化更新,再将更新后的参数下发到每个计算节点。
这样,通过计算节点和参数服务器的分布式协作,可以完成神经网络的SGD方法的训练。PaddlePaddle可以同时支持同步随机梯度下降(SGD)和异步随机梯度下降。
在使用同步SGD训练神经网络时,PaddlePaddle使用同步屏障(barrier),使梯度的提交和参数的更新按照顺序方式执行。在异步SGD中,则并不会等待所有trainer提交梯度才更新参数,这样极大地提高了计算的并行性:参数服务器之间不相互依赖,并行地接收梯度和更新参数,参数服务器也不会等待计算节点全部都提交梯度之后才开始下一步,计算节点之间也不会相互依赖,并行地执行模型的训练。可以看出,虽然异步SGD方式会提高参数更新并行度, 但是并不能保证参数同步更新,在任意时间某一台参数服务器上保存的参数可能比另一台要更新,与同步SGD相比,梯度会有噪声。
## Introduction
In this section, we'll explain how to run distributed training jobs with PaddlePaddle on different types of clusters. The diagram below shows the main architecture of a distributed trainning job:
<img src="https://user-images.githubusercontent.com/13348433/31772146-41523d84-b511-11e7-8a12-a69fd136c283.png" width="500">
- Data shard: training data will be split into multiple partitions, trainers use the partitions of the whole dataset to do the training job.
- Trainer: each trainer reads the data shard, and train the neural network. Then the trainer will upload calculated "gradients" to parameter servers, and wait for parameters to be optimized on the parameter server side. When that finishes, the trainer download optimized parameters and continues its training.
- Parameter server: every parameter server stores part of the whole neural network model data. They will do optimization calculations when gradients are uploaded from trainers, and then send updated parameters to trainers.
PaddlePaddle can support both synchronize stochastic gradient descent (SGD) and asynchronous SGD.
When training with synchronize SGD, PaddlePaddle uses an internal "synchronize barrier" which makes gradients update and parameter download in strict order. On the other hand, asynchronous SGD won't wait for all trainers to finish upload at a single step, this will increase the parallelism of distributed training: parameter servers do not depend on each other, they'll do parameter optimization concurrently. Parameter servers will not wait for trainers, so trainers will also do their work concurrently. But asynchronous SGD will introduce more randomness and noises in the gradient.
在不同集群中运行
================
PaddlePaddle可以使用多种分布式计算平台构建分布式计算任务,包括:
- `Kubernetes <http://kubernetes.io>`_ Google开源的容器集群的调度框架,支持大规模集群生产环境的完整集群方案。
- `OpenMPI <https://www.open-mpi.org>`_ 成熟的高性能并行计算框架。
- `Fabric <http://www.fabfile.org>`_ 集群管理工具。可以使用`Fabric`编写集群任务提交和管理脚本。
对于不同的集群平台,会分别介绍集群作业的启动和停止方法。这些例子都可以在 `cluster_train_v2 <https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/scripts/cluster_train_v2>`_ 找到。
在使用分布式计算平台进行训练时,任务被调度在集群中时,分布式计算平台通常会通过API或者环境变量提供任务运行需要的参数,比如节点的ID、IP和任务节点个数等。
.. toctree::
:maxdepth: 1
fabric_cn.md
openmpi_cn.md
k8s_cn.md
k8s_distributed_cn.md
k8s_aws_cn.md
Use different clusters
======================
PaddlePaddle supports running jobs on several platforms including:
- `Kubernetes <http://kubernetes.io>`_ open-source system for automating deployment, scaling, and management of containerized applications from Google.
- `OpenMPI <https://www.open-mpi.org>`_ Mature high performance parallel computing framework.
- `Fabric <http://www.fabfile.org>`_ A cluster management tool. Write scripts to submit jobs or manage the cluster.
We'll introduce cluster job management on these platforms. The examples can be found under `cluster_train_v2 <https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/scripts/cluster_train_v2>`_ .
These cluster platforms provide API or environment variables for training processes, when the job is dispatched to different nodes. Like node ID, IP or total number of nodes etc.
.. toctree::
:maxdepth: 1
fabric_en.md
openmpi_en.md
k8s_en.md
k8s_aws_en.md
## 环境准备
1. 准备您的计算集群。计算集群通常由一组(几台到几千台规模)的Linux服务器组成。服务器之间可以通过局域网(LAN)联通,每台服务器具有集群中唯一的IP地址(或者可被DNS解析的主机名)。集群中的每台计算机通常被成为一个“节点”。
1. 我们需要在集群的所有节点上安装 PaddlePaddle。 如果要启用GPU,还需要在节点上安装对应的GPU驱动以及CUDA。PaddlePaddle的安装可以参考[build_and_install](http://www.paddlepaddle.org/docs/develop/documentation/zh/getstarted/build_and_install/index_cn.html)的多种安装方式。我们推荐使用[Docker](http://www.paddlepaddle.org/docs/develop/documentation/zh/getstarted/build_and_install/docker_install_cn.html)安装方式来快速安装PaddlePaddle。
安装完成之后,执行下面的命令可以查看已经安装的版本(docker安装方式可以进入docker容器执行:`docker run -it paddlepaddle/paddle:[tag] /bin/bash`):
```bash
$ paddle version
PaddlePaddle 0.10.0, compiled with
with_avx: ON
with_gpu: OFF
with_double: OFF
with_python: ON
with_rdma: OFF
with_timer: OFF
```
## Preparations
1. Prepare your computer cluster. It's normally a bunch of Linux servers connected by LAN. Each server will be assigned a unique IP address. The computers in the cluster can be called "nodes".
2. Install PaddlePaddle on every node. If you are going to take advantage of GPU cards, you'll also need to install proper driver and CUDA libraries. To install PaddlePaddle please read [this build and install](http://www.paddlepaddle.org/docs/develop/documentation/en/getstarted/build_and_install/index_en.html) document. We strongly recommend using [Docker installation](http://www.paddlepaddle.org/docs/develop/documentation/en/getstarted/build_and_install/docker_install_en.html).
After installation, you can check the version by typing the below command (run a docker container if using docker: `docker run -it paddlepaddle/paddle:[tag] /bin/bash`):
```bash
$ paddle version
PaddlePaddle 0.10.0rc, compiled with
with_avx: ON
with_gpu: OFF
with_double: OFF
with_python: ON
with_rdma: OFF
with_timer: OFF
```
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
此差异已折叠。
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