def run():
user_ms = xdl.ModelServer(
"user_graph", user_graph_train, xdl.DataType.float,
xdl.ModelServer.Forward.UniqueCache(xdl.get_task_num()),
xdl.ModelServer.Backward.UniqueCache(xdl.get_task_num()))
xdl.current_env().start_model_server(user_ms)
ad_ms = xdl.ModelServer(
"ad_graph", ad_graph_train, xdl.DataType.float,
xdl.ModelServer.Forward.UniqueCache(xdl.get_task_num()),
xdl.ModelServer.Backward.UniqueCache(xdl.get_task_num()))
xdl.current_env().start_model_server(ad_ms)
batch = reader().read()
user0 = xdl.embedding("user0",
batch["user0"],
xdl.TruncatedNormal(stddev=0.001),
16,
2 * 1024 * 1024,
"sum",
vtype="hash")
user1 = xdl.embedding("user1",
batch["user1"],
xdl.TruncatedNormal(stddev=0.001),
16,
2 * 1024 * 1024,
"sum",
vtype="hash")
ad0 = batch["ad0"]
ad1 = batch["ad1"]
img0 = user_ms(batch["user_img"].ids)
ids0 = xdl.py_func(to_tf_segment_id, [batch["user_img"].segments],
[np.int32])[0]
img1 = ad_ms(batch["ad_img"].ids)
ids1 = xdl.py_func(to_tf_segment_id, [batch["ad_img"].segments],
[np.int32])[0]
label = batch['label']
loss = ams_main(main_model)(user0,
user1,
ad0,
ad1,
label,
ids0,
ids1,
gear_inputs=[img0, img1])
optimizer = xdl.Adam(0.0005).optimize()
run_ops = [loss, optimizer]
sess = xdl.TrainSession([])
while not sess.should_stop():
values = sess.run(run_ops)
if values is not None:
print 'loss: ', values[0]
def train():
batch = reader.read()
sess = xdl.TrainSession()
emb1 = xdl.embedding('emb1', batch['sparse0'], xdl.TruncatedNormal(stddev=0.001), 8, 1024, vtype='hash')
emb2 = xdl.embedding('emb2', batch['sparse1'], xdl.TruncatedNormal(stddev=0.001), 8, 1024, vtype='hash')
loss = model(batch['deep0'], [emb1, emb2], batch['label'])
train_op = xdl.SGD(0.5).optimize()
log_hook = xdl.LoggerHook(loss, "loss:{0}", 10)
sess = xdl.TrainSession(hooks=[log_hook])
while not sess.should_stop():
sess.run(train_op)
def get_xdl_initializer(name="glorot"):
if name == "const":
return xdl.Constant(0.3)
elif name == "glorot":
return xdl.VarianceScaling(scale=1.0,
mode="fan_avg",
distribution="normal",
seed=3)
elif name == "normal":
return xdl.TruncatedNormal(stddev=0.36, seed=3)
def run(is_training, files):
data_io = reader("esmm", files, 2, batch_size, 2, user_fn, ad_fn)
batch = data_io.read()
user_embs = list()
for fn in user_fn:
emb = xdl.embedding('u_' + fn,
batch[fn],
xdl.TruncatedNormal(stddev=0.001),
embed_size,
1000,
'sum',
vtype='hash')
user_embs.append(emb)
ad_embs = list()
for fn in ad_fn:
emb = xdl.embedding('a_' + fn,
batch[fn],
xdl.TruncatedNormal(stddev=0.001),
embed_size,
1000,
'sum',
vtype='hash')
ad_embs.append(emb)
var_list = model(is_training)(ad_embs, user_embs, batch["indicators"][0],
batch["label"])
keys = [
'loss', 'ctr_prop', 'ctcvr_prop', 'cvr_prop', 'ctr_label',
'ctcvr_label', 'cvr_label'
]
run_vars = dict(zip(keys, list(var_list)))
hooks = []
if is_training:
train_op = xdl.Adam(lr).optimize()
hooks = get_collection(READER_HOOKS)
if hooks is None:
hooks = []
if xdl.get_task_index() == 0:
ckpt_hook = xdl.CheckpointHook(1000)
hooks.append(ckpt_hook)
run_vars.update({None: train_op})
if is_debug > 1:
print("=========gradients")
grads = xdl.get_gradients()
grads_keys = grads[''].keys()
grads_keys.sort()
for key in grads_keys:
run_vars.update({"grads {}".format(key): grads[''][key]})
hooks.append(QpsMetricsHook())
log_format = "lstep[%(lstep)s] gstep[%(gstep)s] " \
"lqps[%(lqps)s] gqps[%(gqps)s]"
hooks.append(MetricsPrinterHook(log_format, 100))
ckpt = xdl.get_config("checkpoint", "ckpt")
if ckpt is not None and len(ckpt) > 0:
if int(xdl.get_task_index()) == 0:
from xdl.python.training.saver import Saver
saver = Saver()
print("restore from %s" % ckpt)
saver.restore(ckpt)
else:
time.sleep(120)
sess = xdl.TrainSession(hooks)
if is_training:
itr = 1
ctr_auc = Auc('ctr')
ctcvr_auc = Auc('ctcvr')
cvr_auc = Auc('cvr')
while not sess.should_stop():
print('iter=', itr)
values = sess.run(run_vars.values())
if not values:
continue
value_map = dict(zip(run_vars.keys(), values))
print('loss=', value_map['loss'])
ctr_auc.add(value_map['ctr_prop'], value_map['ctr_label'])
ctcvr_auc.add(value_map['ctcvr_prop'], value_map['ctcvr_label'])
cvr_auc.add_with_filter(value_map['cvr_prop'],
value_map['cvr_label'],
np.where(value_map['ctr_label'] == 1))
itr += 1
ctr_auc.show()
ctcvr_auc.show()
cvr_auc.show()
else:
ctr_test_auc = Auc('ctr')
ctcvr_test_auc = Auc('ctcvr')
cvr_test_auc = Auc('cvr')
for i in xrange(test_batch_num):
print('iter=', i + 1)
values = sess.run(run_vars.values())
value_map = dict(zip(run_vars.keys(), values))
print('test_loss=', value_map['loss'])
ctr_test_auc.add(value_map['ctr_prop'], value_map['ctr_label'])
ctcvr_test_auc.add(value_map['ctcvr_prop'],
value_map['ctcvr_label'])
cvr_test_auc.add_with_filter(value_map['cvr_prop'],
value_map['cvr_label'],
np.where(value_map['ctr_label'] == 1))
ctr_test_auc.show()
ctcvr_test_auc.show()
cvr_test_auc.show()