Skip to content

B
book

PaddlePaddle / book

通知 16
Star 4
Fork 0
B
book
未验证 提交 aaefadae 编写于 作者: L lujun 提交者： GitHub
浏览文件
操作

Merge pull request #647 from junjun315/05-stuff 

update to low level api--05 recommender system
上级 65036c6c 072c8c29
隐藏空白更改
内联 并排
Showing with 356 addition and 227 deletion
05.recommender_system/README.cn.md
浏览文件 @ aaefadae
......@@ -225,15 +225,6 @@ import paddle
import paddle.fluid as fluid
import paddle.fluid.layers as layers
import paddle.fluid.nets as nets
try:
from paddle.fluid.contrib.trainer import *
from paddle.fluid.contrib.inferencer import *
except ImportError:
print(
"In the fluid 1.0, the trainer and inferencer are moving to paddle.fluid.contrib",
file=sys.stderr)
from paddle.fluid.trainer import *
from paddle.fluid.inferencer import *
IS_SPARSE = True
USE_GPU = False
......@@ -414,13 +405,8 @@ test_reader = paddle.batch(
paddle.dataset.movielens.test(), batch_size=BATCH_SIZE)
```
### 构造训练器(trainer)
训练器需要一个训练程序和一个训练优化函数。
```python
trainer = Trainer(
train_func=train_program, place=place, optimizer_func=optimizer_func)
```
### 构造训练过程(trainer)
我们这里构造了一个训练过程，包括训练优化函数。
### 提供数据
......@@ -433,56 +419,92 @@ feed_order = [
]
```
### 事件处理器
回调函数`event_handler`在一个之前定义好的事件发生后会被调用。例如，我们可以在每步训练结束后查看误差。
### 构建训练程序以及测试程序
分别构建训练程序和测试程序，并引入训练优化器。
```python
main_program = fluid.default_main_program()
star_program = fluid.default_startup_program()
[avg_cost, scale_infer] = train_program()
test_program = main_program.clone(for_test=True)
sgd_optimizer = optimizer_func()
sgd_optimizer.minimize(avg_cost)
exe = fluid.Executor(place)
def train_test(program, reader):
count = 0
feed_var_list = [
program.global_block().var(var_name) for var_name in feed_order
]
feeder_test = fluid.DataFeeder(
feed_list=feed_var_list, place=place)
test_exe = fluid.Executor(place)
accumulated = len([avg_cost, scale_infer]) * [0]
for test_data in reader():
avg_cost_np = test_exe.run(program=program,
feed=feeder_test.feed(test_data),
fetch_list=[avg_cost, scale_infer])
accumulated = [x[0] + x[1][0] for x in zip(accumulated, avg_cost_np)]
count += 1
return [x / count for x in accumulated]
```
### 构建训练主循环并开始训练
我们根据上面定义的训练循环数（`PASS_NUM`）和一些别的参数，来进行训练循环，并且每次循环都进行一次测试，当测试结果足够好时退出训练并保存训练好的参数。
```python
# Specify the directory path to save the parameters
params_dirname = "recommender_system.inference.model"
def event_handler(event):
if isinstance(event, EndStepEvent):
test_reader = paddle.batch(
paddle.dataset.movielens.test(), batch_size=BATCH_SIZE)
avg_cost_set = trainer.test(
reader=test_reader, feed_order=feed_order)
# get avg cost
avg_cost = np.array(avg_cost_set).mean()
print("avg_cost: %s" % avg_cost)
if float(avg_cost) < 4: # Change this number to adjust accuracy
trainer.save_params(params_dirname)
trainer.stop()
else:
print('BatchID {0}, Test Loss {1:0.2}'.format(event.epoch + 1,
float(avg_cost)))
if math.isnan(float(avg_cost)):
from paddle.utils.plot import Ploter
test_prompt = "Test cost"
plot_cost = Ploter(test_prompt)
def train_loop():
feed_list = [
main_program.global_block().var(var_name) for var_name in feed_order
]
feeder = fluid.DataFeeder(feed_list, place)
exe.run(star_program)
for pass_id in range(PASS_NUM):
for batch_id, data in enumerate(train_reader()):
# train a mini-batch
outs = exe.run(program=main_program,
feed=feeder.feed(data),
fetch_list=[avg_cost])
out = np.array(outs[0])
avg_cost_set = train_test(test_program, test_reader)
# get test avg_cost
test_avg_cost = np.array(avg_cost_set).mean()
plot_cost.append(test_prompt, batch_id, outs[0])
plot_cost.plot()
print("avg_cost: %s" % test_avg_cost)
if batch_id == 20:
if params_dirname is not None:
fluid.io.save_inference_model(params_dirname, [
"user_id", "gender_id", "age_id", "job_id",
"movie_id", "category_id", "movie_title"
], [scale_infer], exe)
return
else:
print('BatchID {0}, Test Loss {1:0.2}'.format(pass_id + 1,
float(test_avg_cost)))
if math.isnan(float(out[0])):
sys.exit("got NaN loss, training failed.")
```
### 开始训练
最后，我们传入训练循环数（`num_epoch`）和一些别的参数，调用 `trainer.train` 来开始训练。
```python
trainer.train(
num_epochs=1,
event_handler=event_handler,
reader=train_reader,
feed_order=feed_order)
train_loop()
```
## 应用模型
### 构建预测器
传入`inference_program`和`params_dirname`来初始化一个预测器, `params_dirname`用来存放训练过程中的各个参数。
```python
inferencer = Inferencer(
inference_program, param_path=params_dirname, place=place)
```
### 生成测试用输入数据
### 生成测试数据
使用 create_lod_tensor(data, lod, place) 的API来生成细节层次的张量。`data`是一个序列，每个元素是一个索引号的序列。`lod`是细节层次的信息，对应于`data`。比如，data = [[10, 2, 3], [2, 3]] 意味着它包含两个序列，长度分别是3和2。于是相应地 lod = [[3, 2]]，它表明其包含一层细节信息，意味着 `data` 有两个序列，长度分别是3和2。
在这个预测例子中，我们试着预测用户ID为1的用户对于电影'Hunchback of Notre Dame'的评分
......@@ -500,27 +522,42 @@ movie_title = fluid.create_lod_tensor([[1069, 4140, 2923, 710, 988]], [[5]],
place) # 'hunchback','of','notre','dame','the'
```
### 构建预测过程并测试
与训练过程类似，我们需要构建一个预测过程。其中， `params_dirname`是之前用来存放训练过程中的各个参数的地址。
```python
place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
exe = fluid.Executor(place)
inference_scope = fluid.core.Scope()
```
### 测试
现在我们可以进行预测了。我们要提供的`feed_order`应该和训练过程一致。
```python
results = inferencer.infer(
{
'user_id': user_id,
'gender_id': gender_id,
'age_id': age_id,
'job_id': job_id,
'movie_id': movie_id,
'category_id': category_id,
'movie_title': movie_title
},
return_numpy=False)
predict_rating = np.array(results[0])
print("Predict Rating of user id 1 on movie \"" + infer_movie_name + "\" is " + str(predict_rating[0][0]))
print("Actual Rating of user id 1 on movie \"" + infer_movie_name + "\" is 4.")
with fluid.scope_guard(inference_scope):
[inferencer, feed_target_names,
fetch_targets] = fluid.io.load_inference_model(params_dirname, exe)
results = exe.run(inferencer,
feed={
'user_id': user_id,
'gender_id': gender_id,
'age_id': age_id,
'job_id': job_id,
'movie_id': movie_id,
'category_id': category_id,
'movie_title': movie_title
},
fetch_list=fetch_targets,
return_numpy=False)
predict_rating = np.array(results[0])
print("Predict Rating of user id 1 on movie \"" + infer_movie_name +
"\" is " + str(predict_rating[0][0]))
print("Actual Rating of user id 1 on movie \"" + infer_movie_name +
"\" is 4.")
```
## 总结
......
05.recommender_system/index.cn.html
浏览文件 @ aaefadae
......@@ -267,15 +267,6 @@ import paddle
import paddle.fluid as fluid
import paddle.fluid.layers as layers
import paddle.fluid.nets as nets
try:
from paddle.fluid.contrib.trainer import *
from paddle.fluid.contrib.inferencer import *
except ImportError:
print(
"In the fluid 1.0, the trainer and inferencer are moving to paddle.fluid.contrib",
file=sys.stderr)
from paddle.fluid.trainer import *
from paddle.fluid.inferencer import *
IS_SPARSE = True
USE_GPU = False
......@@ -456,13 +447,8 @@ test_reader = paddle.batch(
paddle.dataset.movielens.test(), batch_size=BATCH_SIZE)
```
### 构造训练器(trainer)
训练器需要一个训练程序和一个训练优化函数。
```python
trainer = Trainer(
train_func=train_program, place=place, optimizer_func=optimizer_func)
```
### 构造训练过程(trainer)
我们这里构造了一个训练过程，包括训练优化函数。
### 提供数据
......@@ -475,56 +461,92 @@ feed_order = [
]
```
### 事件处理器
回调函数`event_handler`在一个之前定义好的事件发生后会被调用。例如，我们可以在每步训练结束后查看误差。
### 构建训练程序以及测试程序
分别构建训练程序和测试程序，并引入训练优化器。
```python
main_program = fluid.default_main_program()
star_program = fluid.default_startup_program()
[avg_cost, scale_infer] = train_program()
test_program = main_program.clone(for_test=True)
sgd_optimizer = optimizer_func()
sgd_optimizer.minimize(avg_cost)
exe = fluid.Executor(place)
def train_test(program, reader):
count = 0
feed_var_list = [
program.global_block().var(var_name) for var_name in feed_order
]
feeder_test = fluid.DataFeeder(
feed_list=feed_var_list, place=place)
test_exe = fluid.Executor(place)
accumulated = len([avg_cost, scale_infer]) * [0]
for test_data in reader():
avg_cost_np = test_exe.run(program=program,
feed=feeder_test.feed(test_data),
fetch_list=[avg_cost, scale_infer])
accumulated = [x[0] + x[1][0] for x in zip(accumulated, avg_cost_np)]
count += 1
return [x / count for x in accumulated]
```
### 构建训练主循环并开始训练
我们根据上面定义的训练循环数（`PASS_NUM`）和一些别的参数，来进行训练循环，并且每次循环都进行一次测试，当测试结果足够好时退出训练并保存训练好的参数。
```python
# Specify the directory path to save the parameters
params_dirname = "recommender_system.inference.model"
def event_handler(event):
if isinstance(event, EndStepEvent):
test_reader = paddle.batch(
paddle.dataset.movielens.test(), batch_size=BATCH_SIZE)
avg_cost_set = trainer.test(
reader=test_reader, feed_order=feed_order)
# get avg cost
avg_cost = np.array(avg_cost_set).mean()
print("avg_cost: %s" % avg_cost)
if float(avg_cost) < 4: # Change this number to adjust accuracy
trainer.save_params(params_dirname)
trainer.stop()
else:
print('BatchID {0}, Test Loss {1:0.2}'.format(event.epoch + 1,
float(avg_cost)))
if math.isnan(float(avg_cost)):
from paddle.utils.plot import Ploter
test_prompt = "Test cost"
plot_cost = Ploter(test_prompt)
def train_loop():
feed_list = [
main_program.global_block().var(var_name) for var_name in feed_order
]
feeder = fluid.DataFeeder(feed_list, place)
exe.run(star_program)
for pass_id in range(PASS_NUM):
for batch_id, data in enumerate(train_reader()):
# train a mini-batch
outs = exe.run(program=main_program,
feed=feeder.feed(data),
fetch_list=[avg_cost])
out = np.array(outs[0])
avg_cost_set = train_test(test_program, test_reader)
# get test avg_cost
test_avg_cost = np.array(avg_cost_set).mean()
plot_cost.append(test_prompt, batch_id, outs[0])
plot_cost.plot()
print("avg_cost: %s" % test_avg_cost)
if batch_id == 20:
if params_dirname is not None:
fluid.io.save_inference_model(params_dirname, [
"user_id", "gender_id", "age_id", "job_id",
"movie_id", "category_id", "movie_title"
], [scale_infer], exe)
return
else:
print('BatchID {0}, Test Loss {1:0.2}'.format(pass_id + 1,
float(test_avg_cost)))
if math.isnan(float(out[0])):
sys.exit("got NaN loss, training failed.")
```
### 开始训练
最后，我们传入训练循环数（`num_epoch`）和一些别的参数，调用 `trainer.train` 来开始训练。
```python
trainer.train(
num_epochs=1,
event_handler=event_handler,
reader=train_reader,
feed_order=feed_order)
train_loop()
```
## 应用模型
### 构建预测器
传入`inference_program`和`params_dirname`来初始化一个预测器, `params_dirname`用来存放训练过程中的各个参数。
```python
inferencer = Inferencer(
inference_program, param_path=params_dirname, place=place)
```
### 生成测试用输入数据
### 生成测试数据
使用 create_lod_tensor(data, lod, place) 的API来生成细节层次的张量。`data`是一个序列，每个元素是一个索引号的序列。`lod`是细节层次的信息，对应于`data`。比如，data = [[10, 2, 3], [2, 3]] 意味着它包含两个序列，长度分别是3和2。于是相应地 lod = [[3, 2]]，它表明其包含一层细节信息，意味着 `data` 有两个序列，长度分别是3和2。
在这个预测例子中，我们试着预测用户ID为1的用户对于电影'Hunchback of Notre Dame'的评分
......@@ -542,27 +564,42 @@ movie_title = fluid.create_lod_tensor([[1069, 4140, 2923, 710, 988]], [[5]],
place) # 'hunchback','of','notre','dame','the'
```
### 构建预测过程并测试
与训练过程类似，我们需要构建一个预测过程。其中， `params_dirname`是之前用来存放训练过程中的各个参数的地址。
```python
place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
exe = fluid.Executor(place)
inference_scope = fluid.core.Scope()
```
### 测试
现在我们可以进行预测了。我们要提供的`feed_order`应该和训练过程一致。
```python
results = inferencer.infer(
{
'user_id': user_id,
'gender_id': gender_id,
'age_id': age_id,
'job_id': job_id,
'movie_id': movie_id,
'category_id': category_id,
'movie_title': movie_title
},
return_numpy=False)
predict_rating = np.array(results[0])
print("Predict Rating of user id 1 on movie \"" + infer_movie_name + "\" is " + str(predict_rating[0][0]))
print("Actual Rating of user id 1 on movie \"" + infer_movie_name + "\" is 4.")
with fluid.scope_guard(inference_scope):
[inferencer, feed_target_names,
fetch_targets] = fluid.io.load_inference_model(params_dirname, exe)
results = exe.run(inferencer,
feed={
'user_id': user_id,
'gender_id': gender_id,
'age_id': age_id,
'job_id': job_id,
'movie_id': movie_id,
'category_id': category_id,
'movie_title': movie_title
},
fetch_list=fetch_targets,
return_numpy=False)
predict_rating = np.array(results[0])
print("Predict Rating of user id 1 on movie \"" + infer_movie_name +
"\" is " + str(predict_rating[0][0]))
print("Actual Rating of user id 1 on movie \"" + infer_movie_name +
"\" is 4.")
```
## 总结
......
05.recommender_system/train.py
浏览文件 @ aaefadae
......@@ -20,19 +20,11 @@ import paddle
import paddle.fluid as fluid
import paddle.fluid.layers as layers
import paddle.fluid.nets as nets
try:
from paddle.fluid.contrib.trainer import *
from paddle.fluid.contrib.inferencer import *
except ImportError:
print(
"In the fluid 1.0, the trainer and inferencer are moving to paddle.fluid.contrib",
file=sys.stderr)
from paddle.fluid.trainer import *
from paddle.fluid.inferencer import *
IS_SPARSE = True
USE_GPU = False
BATCH_SIZE = 256
PASS_NUM = 100
def get_usr_combined_features():
......@@ -148,71 +140,101 @@ def inference_program():
inference = layers.cos_sim(X=usr_combined_features, Y=mov_combined_features)
scale_infer = layers.scale(x=inference, scale=5.0)
return scale_infer
def train_program():
scale_infer = inference_program()
label = layers.data(name='score', shape=[1], dtype='float32')
square_cost = layers.square_error_cost(input=scale_infer, label=label)
avg_cost = layers.mean(square_cost)
return [avg_cost, scale_infer]
return scale_infer, avg_cost
def optimizer_func():
return fluid.optimizer.SGD(learning_rate=0.2)
def train(use_cuda, train_program, params_dirname):
def train(use_cuda, params_dirname):
place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
trainer = Trainer(
train_func=train_program, place=place, optimizer_func=optimizer_func)
train_reader = paddle.batch(
paddle.reader.shuffle(paddle.dataset.movielens.train(), buf_size=8192),
batch_size=BATCH_SIZE)
test_reader = paddle.batch(
paddle.dataset.movielens.test(), batch_size=BATCH_SIZE)
feed_order = [
'user_id', 'gender_id', 'age_id', 'job_id', 'movie_id', 'category_id',
'movie_title', 'score'
]
def event_handler(event):
if isinstance(event, EndStepEvent):
test_reader = paddle.batch(
paddle.dataset.movielens.test(), batch_size=BATCH_SIZE)
avg_cost_set = trainer.test(
reader=test_reader, feed_order=feed_order)
# get avg cost
avg_cost = np.array(avg_cost_set).mean()
print("avg_cost: %s" % avg_cost)
if float(avg_cost) < 4: # Change this number to adjust accuracy
trainer.save_params(params_dirname)
trainer.stop()
else:
print('BatchID {0}, Test Loss {1:0.2}'.format(event.epoch + 1,
float(avg_cost)))
if math.isnan(float(avg_cost)):
main_program = fluid.default_main_program()
star_program = fluid.default_startup_program()
scale_infer, avg_cost = inference_program()
test_program = main_program.clone(for_test=True)
sgd_optimizer = optimizer_func()
sgd_optimizer.minimize(avg_cost)
exe = fluid.Executor(place)
def train_test(program, reader):
count = 0
feed_var_list = [
program.global_block().var(var_name) for var_name in feed_order
]
feeder_test = fluid.DataFeeder(feed_list=feed_var_list, place=place)
test_exe = fluid.Executor(place)
accumulated = len([avg_cost, scale_infer]) * [0]
for test_data in reader():
avg_cost_np = test_exe.run(
program=program,
feed=feeder_test.feed(test_data),
fetch_list=[avg_cost, scale_infer])
accumulated = [
x[0] + x[1][0] for x in zip(accumulated, avg_cost_np)
]
count += 1
return [x / count for x in accumulated]
def train_loop():
feed_list = [
main_program.global_block().var(var_name) for var_name in feed_order
]
feeder = fluid.DataFeeder(feed_list, place)
exe.run(star_program)
for pass_id in range(PASS_NUM):
for batch_id, data in enumerate(train_reader()):
# train a mini-batch
outs = exe.run(
program=main_program,
feed=feeder.feed(data),
fetch_list=[avg_cost])
out = np.array(outs[0])
avg_cost_set = train_test(test_program, test_reader)
# get test avg_cost
test_avg_cost = np.array(avg_cost_set).mean()
print("avg_cost: %s" % test_avg_cost)
# if test_avg_cost < 4.0: # Change this number to adjust accuracy
if batch_id == 20:
if params_dirname is not None:
fluid.io.save_inference_model(params_dirname, [
"user_id", "gender_id", "age_id", "job_id",
"movie_id", "category_id", "movie_title"
], [scale_infer], exe)
return
else:
print('BatchID {0}, Test Loss {1:0.2}'.format(
pass_id + 1, float(test_avg_cost)))
if math.isnan(float(out[0])):
sys.exit("got NaN loss, training failed.")
train_reader = paddle.batch(
paddle.reader.shuffle(paddle.dataset.movielens.train(), buf_size=8192),
batch_size=BATCH_SIZE)
trainer.train(
num_epochs=1,
event_handler=event_handler,
reader=train_reader,
feed_order=feed_order)
train_loop()
def infer(use_cuda, inference_program, params_dirname):
def infer(use_cuda, params_dirname):
place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
inferencer = Inferencer(
inference_program, param_path=params_dirname, place=place)
# Use the first data from paddle.dataset.movielens.test() as input.
# Use create_lod_tensor(data, lod, place) API to generate LoD Tensor,
......@@ -225,46 +247,79 @@ def infer(use_cuda, inference_program, params_dirname):
infer_movie_id = 783
infer_movie_name = paddle.dataset.movielens.movie_info()[
infer_movie_id].title
user_id = fluid.create_lod_tensor([[1]], [[1]], place)
gender_id = fluid.create_lod_tensor([[1]], [[1]], place)
age_id = fluid.create_lod_tensor([[0]], [[1]], place)
job_id = fluid.create_lod_tensor([[10]], [[1]], place)
movie_id = fluid.create_lod_tensor([[783]], [[1]], place)
category_id = fluid.create_lod_tensor([[10, 8, 9]], [[3]], place)
movie_title = fluid.create_lod_tensor([[1069, 4140, 2923, 710, 988]], [[5]],
place)
results = inferencer.infer(
{
'user_id': user_id,
'gender_id': gender_id,
'age_id': age_id,
'job_id': job_id,
'movie_id': movie_id,
'category_id': category_id,
'movie_title': movie_title
},
return_numpy=False)
predict_rating = np.array(results[0])
print("Predict Rating of user id 1 on movie \"" + infer_movie_name +
"\" is " + str(predict_rating[0][0]))
print("Actual Rating of user id 1 on movie \"" + infer_movie_name +
"\" is 4.")
exe = fluid.Executor(place)
inference_scope = fluid.core.Scope()
with fluid.scope_guard(inference_scope):
# Use fluid.io.load_inference_model to obtain the inference program desc,
# the feed_target_names (the names of variables that will be feeded
# data using feed operators), and the fetch_targets (variables that
# we want to obtain data from using fetch operators).
[inferencer, feed_target_names,
fetch_targets] = fluid.io.load_inference_model(params_dirname, exe)
# Use the first data from paddle.dataset.movielens.test() as input
assert feed_target_names[0] == "user_id"
# Use create_lod_tensor(data, recursive_sequence_lengths, place) API
# to generate LoD Tensor where `data` is a list of sequences of index
# numbers, `recursive_sequence_lengths` is the length-based level of detail
# (lod) info associated with `data`.
# For example, data = [[10, 2, 3], [2, 3]] means that it contains
# two sequences of indexes, of length 3 and 2, respectively.
# Correspondingly, recursive_sequence_lengths = [[3, 2]] contains one
# level of detail info, indicating that `data` consists of two sequences
# of length 3 and 2, respectively.
user_id = fluid.create_lod_tensor([[1]], [[1]], place)
assert feed_target_names[1] == "gender_id"
gender_id = fluid.create_lod_tensor([[1]], [[1]], place)
assert feed_target_names[2] == "age_id"
age_id = fluid.create_lod_tensor([[0]], [[1]], place)
assert feed_target_names[3] == "job_id"
job_id = fluid.create_lod_tensor([[10]], [[1]], place)
assert feed_target_names[4] == "movie_id"
movie_id = fluid.create_lod_tensor([[783]], [[1]], place)
assert feed_target_names[5] == "category_id"
category_id = fluid.create_lod_tensor([[10, 8, 9]], [[3]], place)
assert feed_target_names[6] == "movie_title"
movie_title = fluid.create_lod_tensor([[1069, 4140, 2923, 710, 988]],
[[5]], place)
# Construct feed as a dictionary of {feed_target_name: feed_target_data}
# and results will contain a list of data corresponding to fetch_targets.
results = exe.run(
inferencer,
feed={
feed_target_names[0]: user_id,
feed_target_names[1]: gender_id,
feed_target_names[2]: age_id,
feed_target_names[3]: job_id,
feed_target_names[4]: movie_id,
feed_target_names[5]: category_id,
feed_target_names[6]: movie_title
},
fetch_list=fetch_targets,
return_numpy=False)
predict_rating = np.array(results[0])
print("Predict Rating of user id 1 on movie \"" + infer_movie_name +
"\" is " + str(predict_rating[0][0]))
print("Actual Rating of user id 1 on movie \"" + infer_movie_name +
"\" is 4.")
def main(use_cuda):
if use_cuda and not fluid.core.is_compiled_with_cuda():
return
params_dirname = "recommender_system.inference.model"
train(
use_cuda=use_cuda,
train_program=train_program,
params_dirname=params_dirname)
infer(
use_cuda=use_cuda,
inference_program=inference_program,
params_dirname=params_dirname)
train(use_cuda=use_cuda, params_dirname=params_dirname)
infer(use_cuda=use_cuda, params_dirname=params_dirname)
if __name__ == '__main__':
......
Markdown is supported
0% 或 .
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 或