def main(argv):
max_iter = 100
matrix = np.random.random((1000, 1000))
jobs_def = [
{
"name": "ps",
"num": 2
},
{
"name": "worker",
"num": 2
},
]
mesos_master = argv[1]
nmf = NMF(matrix, 200)
with cluster(jobs_def, master=mesos_master) as c:
with tf.Session(c.targets['/job:worker/task:1']) as session:
session.run(tf.global_variables_initializer())
for i in range(max_iter):
mat_w, mat_h, loss = nmf.run(session)
print("loss#%d: %s" % (i, loss))
err = np.power(matrix - np.matmul(mat_w, mat_h), 2)
print("err mean: %s" % err.mean())
print("loss: %s" % loss)
def main(argv):
max_iter = 100
matrix = np.random.random((1000, 1000))
jobs_def = [
{
"name": "ps",
"num": 2
},
{
"name": "worker",
"num": 2
},
]
mesos_master = argv[1]
with cluster(jobs_def, master=mesos_master) as targets:
with tf.Session(targets['/job:worker/task:1']) as session:
nmf = NMF(session, matrix, 200)
for i in range(max_iter):
mat_w, mat_h, loss = nmf.run()
print "loss#%d: %s" % (i, loss)
err = np.power(matrix - np.matmul(mat_w, mat_h), 2)
print "err mean: %s" % err.mean()
print "loss: %s" % loss
def main(argv):
jobs_def = [
{
"name": "ps",
"num": 2
},
{
"name": "worker",
"num": 2
},
]
mesos_master = sys.argv[1]
with cluster(jobs_def, master=mesos_master, quiet=False) as targets:
with tf.device('/job:ps/task:0'):
a = tf.constant(10)
with tf.device('/job:ps/task:1'):
b = tf.constant(32)
with tf.device("/job:worker/task:1"):
op = a + b
with tf.Session(targets['/job:worker/task:0']) as sess:
print(sess.run(op))
def main(argv):
jobs_def = [
{
"name": "ps",
"num": 2
},
{
"name": "worker",
"num": 2
},
]
mesos_master = sys.argv[1]
with cluster(jobs_def, master=mesos_master, quiet=False) as c:
with tf.device('/job:ps/task:0'):
a = tf.constant(10)
with tf.device('/job:ps/task:1'):
b = tf.constant(32)
with tf.device("/job:worker/task:1"):
op = a + b
with tf.Session(c.targets['/job:worker/task:0']) as sess:
print(sess.run(op))
jobs_def = [
{
"name": "ps",
"num": nserver
},
{
"name": "worker",
"num": nworker,
"gpus": args.worker_gpus,
},
]
_lock = RLock()
mnist = read_data_sets("MNIST_data/", one_hot=True)
with cluster(jobs_def, master=master, quiet=False, **extra_kw) as c:
graph = tf.Graph()
with graph.as_default():
with tf.device(tf.train.replica_device_setter(ps_tasks=nserver)):
W = tf.Variable(tf.zeros([784, 10]))
b = tf.Variable(tf.zeros([10]))
global_step = tf.Variable(0)
x = tf.placeholder(tf.float32, [None, 784])
y = tf.nn.softmax(tf.matmul(x, W) + b)
y_ = tf.placeholder(tf.float32, [None, 10])
cross_entropy = -tf.reduce_sum(y_ * tf.log(y))
steps = []
for i in range(nworker):
with tf.device('/job:worker/task:%d' % i):
steps.append(
jobs_def = [
{
"name": "ps",
"num": nserver
},
{
"name": "worker",
"num": nworker,
"gpus": args.worker_gpus,
},
]
_lock = RLock()
mnist = read_data_sets("MNIST_data/", one_hot=True)
with cluster(jobs_def, master=master, quiet=False) as targets:
graph = tf.Graph()
with graph.as_default():
with tf.device(tf.train.replica_device_setter(ps_tasks=nserver)):
W = tf.Variable(tf.zeros([784, 10]))
b = tf.Variable(tf.zeros([10]))
global_step = tf.Variable(0)
x = tf.placeholder(tf.float32, [None, 784])
y = tf.nn.softmax(tf.matmul(x, W) + b)
y_ = tf.placeholder(tf.float32, [None, 10])
cross_entropy = -tf.reduce_sum(y_*tf.log(y))
steps = []
for i in range(nworker):
with tf.device('/job:worker/task:%d' % i):
steps.append(tf.train.GradientDescentOptimizer(0.005).minimize(cross_entropy, global_step=global_step))
jobs_def = [
{
"name": "ps",
"num": nserver
},
{
"name": "worker",
"num": nworker,
"gpus": args.worker_gpus,
},
]
_lock = RLock()
mnist = read_data_sets("MNIST_data/", one_hot=True)
with cluster(jobs_def, master=master, quiet=False) as targets:
graph = tf.Graph()
with graph.as_default():
with tf.device(tf.train.replica_device_setter(ps_tasks=nserver)):
W = tf.Variable(tf.zeros([784, 10]))
b = tf.Variable(tf.zeros([10]))
global_step = tf.Variable(0)
x = tf.placeholder(tf.float32, [None, 784])
y = tf.nn.softmax(tf.matmul(x, W) + b)
y_ = tf.placeholder(tf.float32, [None, 10])
cross_entropy = -tf.reduce_sum(y_ * tf.log(y))
steps = []
for i in xrange(nworker):
with tf.device('/job:worker/task:%d' % i):
steps.append(
