mxnet系列-Linear Regression

(1)配置mxnet开发环境

最方便的方式是使用pycharm开发。1. 安装pycharm。 2.打开pycharm的file/default-setting点击安装 mxnet包。

(2)mxnet初试

import mxnet as mx
import numpy as np

import logging
logging.getLogger().setLevel(logging.DEBUG)

#preparing the data
#train_data
train_data = np.random.uniform(0,1,[100,2])
train_label = np.array([train_data[i][0] + 2 * train_data[i][1] for i in range(100)])
batch_size = 1

#Evaluation Data
eval_data = np.array([[7,2],[6,10],[12,2]])
eval_label = np.array([11,26,16])

train_iter = mx.io.NDArrayIter(train_data,train_label, batch_size, shuffle=True,label_name='lin_reg_label')
eval_iter = mx.io.NDArrayIter(eval_data, eval_label, batch_size, shuffle=False)

#define model
X = mx.sym.Variable('data')
Y = mx.sym.Variable('lin_reg_label')
fully_connected_layer  = mx.sym.FullyConnected(data=X, name='fc1', num_hidden = 1)
lro = mx.sym.LinearRegressionOutput(data=fully_connected_layer, label=Y, name="lro")

model = mx.mod.Module(
    symbol = lro ,
    data_names=['data'],
    label_names = ['lin_reg_label']# network structure
)

#mx.viz.plot_network(symbol=lro).view()
model.fit(train_iter, eval_iter,
            optimizer_params={'learning_rate':0.005, 'momentum': 0.9},
            num_epoch=20,
            eval_metric='mse',
            batch_end_callback = mx.callback.Speedometer(batch_size, 2))


print model.predict(eval_iter).asnumpy()

代码流程

1. 新建训练数据 mx.io.NDArrayIter
   其参数为，默认data_name = ‘data’, label_name = ‘softmax_name’，

    def __init__(self, data, label=None, batch_size=1, shuffle=False,
                 last_batch_handle='pad', data_name='data',
                 label_name='softmax_label'):

X = mx.sym.Variable('data') 对应输入数据的data
Y = mx.sym.Variable('lin_reg_label') 输入数据 lin_reg_label

2.建立模型 mx.mod.Module， 其参数为

    def __init__(self, symbol, data_names=('data',), label_names=('softmax_label',),
                 logger=logging, context=ctx.cpu(), work_load_list=None,
                 fixed_param_names=None, state_names=None, group2ctxs=None,
                 compression_params=None):

3.开始训练 model.fit

    def fit(self, train_data, eval_data=None, eval_metric='acc',
            epoch_end_callback=None, batch_end_callback=None, kvstore='local',
            optimizer='sgd', optimizer_params=(('learning_rate', 0.01),),
            eval_end_callback=None,
            eval_batch_end_callback=None, initializer=Uniform(0.01),
            arg_params=None, aux_params=None, allow_missing=False,
            force_rebind=False, force_init=False, begin_epoch=0, num_epoch=None,
            validation_metric=None, monitor=None):