modelCPMWeight.py

#import caffe
#from caffe.proto import caffe_pb2
#use_caffe = True
import re
from google.protobuf import text_format
import mxnet as mx
from generateLabelCPMWeight import *

numofparts = 15
numoflinks = 13

def CPMModel():
   
    data = mx.symbol.Variable(name='data')
    ## heat map of human parts
    heatmaplabel = mx.sym.Variable("heatmaplabel")
    ## part affinity graph
    partaffinityglabel = mx.sym.Variable('partaffinityglabel')

    heatweight = mx.sym.Variable('heatweight')
    
    vecweight = mx.sym.Variable('vecweight')
    
    conv1_1 = mx.symbol.Convolution(name='conv1_1', data=data , num_filter=64, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu1_1 = mx.symbol.Activation(name='relu1_1', data=conv1_1 , act_type='relu')
    conv1_2 = mx.symbol.Convolution(name='conv1_2', data=relu1_1 , num_filter=64, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu1_2 = mx.symbol.Activation(name='relu1_2', data=conv1_2 , act_type='relu')
    pool1_stage1 = mx.symbol.Pooling(name='pool1_stage1', data=relu1_2 , pooling_convention='full', pad=(0,0), kernel=(2,2), stride=(2,2), pool_type='max')
    conv2_1 = mx.symbol.Convolution(name='conv2_1', data=pool1_stage1 , num_filter=128, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu2_1 = mx.symbol.Activation(name='relu2_1', data=conv2_1 , act_type='relu')
    conv2_2 = mx.symbol.Convolution(name='conv2_2', data=relu2_1 , num_filter=128, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu2_2 = mx.symbol.Activation(name='relu2_2', data=conv2_2 , act_type='relu')
    pool2_stage1 = mx.symbol.Pooling(name='pool2_stage1', data=relu2_2 , pooling_convention='full', pad=(0,0), kernel=(2,2), stride=(2,2), pool_type='max')
    conv3_1 = mx.symbol.Convolution(name='conv3_1', data=pool2_stage1 , num_filter=256, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu3_1 = mx.symbol.Activation(name='relu3_1', data=conv3_1 , act_type='relu')
    conv3_2 = mx.symbol.Convolution(name='conv3_2', data=relu3_1 , num_filter=256, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu3_2 = mx.symbol.Activation(name='relu3_2', data=conv3_2 , act_type='relu')
    conv3_3 = mx.symbol.Convolution(name='conv3_3', data=relu3_2 , num_filter=256, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu3_3 = mx.symbol.Activation(name='relu3_3', data=conv3_3 , act_type='relu')
    conv3_4 = mx.symbol.Convolution(name='conv3_4', data=relu3_3 , num_filter=256, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu3_4 = mx.symbol.Activation(name='relu3_4', data=conv3_4 , act_type='relu')
    pool3_stage1 = mx.symbol.Pooling(name='pool3_stage1', data=relu3_4 , pooling_convention='full', pad=(0,0), kernel=(2,2), stride=(2,2), pool_type='max')
    conv4_1 = mx.symbol.Convolution(name='conv4_1', data=pool3_stage1 , num_filter=512, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu4_1 = mx.symbol.Activation(name='relu4_1', data=conv4_1 , act_type='relu')
    conv4_2 = mx.symbol.Convolution(name='conv4_2', data=relu4_1 , num_filter=512, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu4_2 = mx.symbol.Activation(name='relu4_2', data=conv4_2 , act_type='relu')
    conv4_3_CPM = mx.symbol.Convolution(name='conv4_3_CPM', data=relu4_2 , num_filter=256, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu4_3_CPM = mx.symbol.Activation(name='relu4_3_CPM', data=conv4_3_CPM , act_type='relu')
    conv4_4_CPM = mx.symbol.Convolution(name='conv4_4_CPM', data=relu4_3_CPM , num_filter=128, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu4_4_CPM = mx.symbol.Activation(name='relu4_4_CPM', data=conv4_4_CPM , act_type='relu')
    conv5_1_CPM_L1 = mx.symbol.Convolution(name='conv5_1_CPM_L1', data=relu4_4_CPM , num_filter=128, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu5_1_CPM_L1 = mx.symbol.Activation(name='relu5_1_CPM_L1', data=conv5_1_CPM_L1 , act_type='relu')
    conv5_1_CPM_L2 = mx.symbol.Convolution(name='conv5_1_CPM_L2', data=relu4_4_CPM , num_filter=128, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu5_1_CPM_L2 = mx.symbol.Activation(name='relu5_1_CPM_L2', data=conv5_1_CPM_L2 , act_type='relu')
    conv5_2_CPM_L1 = mx.symbol.Convolution(name='conv5_2_CPM_L1', data=relu5_1_CPM_L1 , num_filter=128, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu5_2_CPM_L1 = mx.symbol.Activation(name='relu5_2_CPM_L1', data=conv5_2_CPM_L1 , act_type='relu')
    conv5_2_CPM_L2 = mx.symbol.Convolution(name='conv5_2_CPM_L2', data=relu5_1_CPM_L2 , num_filter=128, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu5_2_CPM_L2 = mx.symbol.Activation(name='relu5_2_CPM_L2', data=conv5_2_CPM_L2 , act_type='relu')
    conv5_3_CPM_L1 = mx.symbol.Convolution(name='conv5_3_CPM_L1', data=relu5_2_CPM_L1 , num_filter=128, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu5_3_CPM_L1 = mx.symbol.Activation(name='relu5_3_CPM_L1', data=conv5_3_CPM_L1 , act_type='relu')
    conv5_3_CPM_L2 = mx.symbol.Convolution(name='conv5_3_CPM_L2', data=relu5_2_CPM_L2 , num_filter=128, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu5_3_CPM_L2 = mx.symbol.Activation(name='relu5_3_CPM_L2', data=conv5_3_CPM_L2 , act_type='relu')
    conv5_4_CPM_L1 = mx.symbol.Convolution(name='conv5_4_CPM_L1', data=relu5_3_CPM_L1 , num_filter=512, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    relu5_4_CPM_L1 = mx.symbol.Activation(name='relu5_4_CPM_L1', data=conv5_4_CPM_L1 , act_type='relu')
    conv5_4_CPM_L2 = mx.symbol.Convolution(name='conv5_4_CPM_L2', data=relu5_3_CPM_L2 , num_filter=512, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    relu5_4_CPM_L2 = mx.symbol.Activation(name='relu5_4_CPM_L2', data=conv5_4_CPM_L2 , act_type='relu')
    conv5_5_CPM_L1 = mx.symbol.Convolution(name='conv5_5_CPM_L1', data=relu5_4_CPM_L1 , num_filter=numoflinks*2, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    conv5_5_CPM_L2 = mx.symbol.Convolution(name='conv5_5_CPM_L2', data=relu5_4_CPM_L2 , num_filter=numofparts, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    concat_stage2 = mx.symbol.Concat(name='concat_stage2', *[conv5_5_CPM_L1,conv5_5_CPM_L2,relu4_4_CPM] )
    Mconv1_stage2_L1 = mx.symbol.Convolution(name='Mconv1_stage2_L1', data=concat_stage2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu1_stage2_L1 = mx.symbol.Activation(name='Mrelu1_stage2_L1', data=Mconv1_stage2_L1 , act_type='relu')
    Mconv1_stage2_L2 = mx.symbol.Convolution(name='Mconv1_stage2_L2', data=concat_stage2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu1_stage2_L2 = mx.symbol.Activation(name='Mrelu1_stage2_L2', data=Mconv1_stage2_L2 , act_type='relu')
    Mconv2_stage2_L1 = mx.symbol.Convolution(name='Mconv2_stage2_L1', data=Mrelu1_stage2_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu2_stage2_L1 = mx.symbol.Activation(name='Mrelu2_stage2_L1', data=Mconv2_stage2_L1 , act_type='relu')
    Mconv2_stage2_L2 = mx.symbol.Convolution(name='Mconv2_stage2_L2', data=Mrelu1_stage2_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu2_stage2_L2 = mx.symbol.Activation(name='Mrelu2_stage2_L2', data=Mconv2_stage2_L2 , act_type='relu')
    Mconv3_stage2_L1 = mx.symbol.Convolution(name='Mconv3_stage2_L1', data=Mrelu2_stage2_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu3_stage2_L1 = mx.symbol.Activation(name='Mrelu3_stage2_L1', data=Mconv3_stage2_L1 , act_type='relu')
    Mconv3_stage2_L2 = mx.symbol.Convolution(name='Mconv3_stage2_L2', data=Mrelu2_stage2_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu3_stage2_L2 = mx.symbol.Activation(name='Mrelu3_stage2_L2', data=Mconv3_stage2_L2 , act_type='relu')
    Mconv4_stage2_L1 = mx.symbol.Convolution(name='Mconv4_stage2_L1', data=Mrelu3_stage2_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu4_stage2_L1 = mx.symbol.Activation(name='Mrelu4_stage2_L1', data=Mconv4_stage2_L1 , act_type='relu')
    Mconv4_stage2_L2 = mx.symbol.Convolution(name='Mconv4_stage2_L2', data=Mrelu3_stage2_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu4_stage2_L2 = mx.symbol.Activation(name='Mrelu4_stage2_L2', data=Mconv4_stage2_L2 , act_type='relu')
    Mconv5_stage2_L1 = mx.symbol.Convolution(name='Mconv5_stage2_L1', data=Mrelu4_stage2_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu5_stage2_L1 = mx.symbol.Activation(name='Mrelu5_stage2_L1', data=Mconv5_stage2_L1 , act_type='relu')
    Mconv5_stage2_L2 = mx.symbol.Convolution(name='Mconv5_stage2_L2', data=Mrelu4_stage2_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu5_stage2_L2 = mx.symbol.Activation(name='Mrelu5_stage2_L2', data=Mconv5_stage2_L2 , act_type='relu')
    Mconv6_stage2_L1 = mx.symbol.Convolution(name='Mconv6_stage2_L1', data=Mrelu5_stage2_L1 , num_filter=128, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mrelu6_stage2_L1 = mx.symbol.Activation(name='Mrelu6_stage2_L1', data=Mconv6_stage2_L1 , act_type='relu')
    Mconv6_stage2_L2 = mx.symbol.Convolution(name='Mconv6_stage2_L2', data=Mrelu5_stage2_L2 , num_filter=128, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mrelu6_stage2_L2 = mx.symbol.Activation(name='Mrelu6_stage2_L2', data=Mconv6_stage2_L2 , act_type='relu')
    Mconv7_stage2_L1 = mx.symbol.Convolution(name='Mconv7_stage2_L1', data=Mrelu6_stage2_L1 , num_filter=numoflinks*2, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mconv7_stage2_L2 = mx.symbol.Convolution(name='Mconv7_stage2_L2', data=Mrelu6_stage2_L2 , num_filter=numofparts, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    concat_stage3 = mx.symbol.Concat(name='concat_stage3', *[Mconv7_stage2_L1,Mconv7_stage2_L2,relu4_4_CPM] )
    Mconv1_stage3_L1 = mx.symbol.Convolution(name='Mconv1_stage3_L1', data=concat_stage3 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu1_stage3_L1 = mx.symbol.Activation(name='Mrelu1_stage3_L1', data=Mconv1_stage3_L1 , act_type='relu')
    Mconv1_stage3_L2 = mx.symbol.Convolution(name='Mconv1_stage3_L2', data=concat_stage3 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu1_stage3_L2 = mx.symbol.Activation(name='Mrelu1_stage3_L2', data=Mconv1_stage3_L2 , act_type='relu')
    Mconv2_stage3_L1 = mx.symbol.Convolution(name='Mconv2_stage3_L1', data=Mrelu1_stage3_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu2_stage3_L1 = mx.symbol.Activation(name='Mrelu2_stage3_L1', data=Mconv2_stage3_L1 , act_type='relu')
    Mconv2_stage3_L2 = mx.symbol.Convolution(name='Mconv2_stage3_L2', data=Mrelu1_stage3_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu2_stage3_L2 = mx.symbol.Activation(name='Mrelu2_stage3_L2', data=Mconv2_stage3_L2 , act_type='relu')
    Mconv3_stage3_L1 = mx.symbol.Convolution(name='Mconv3_stage3_L1', data=Mrelu2_stage3_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu3_stage3_L1 = mx.symbol.Activation(name='Mrelu3_stage3_L1', data=Mconv3_stage3_L1 , act_type='relu')
    Mconv3_stage3_L2 = mx.symbol.Convolution(name='Mconv3_stage3_L2', data=Mrelu2_stage3_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu3_stage3_L2 = mx.symbol.Activation(name='Mrelu3_stage3_L2', data=Mconv3_stage3_L2 , act_type='relu')
    Mconv4_stage3_L1 = mx.symbol.Convolution(name='Mconv4_stage3_L1', data=Mrelu3_stage3_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu4_stage3_L1 = mx.symbol.Activation(name='Mrelu4_stage3_L1', data=Mconv4_stage3_L1 , act_type='relu')
    Mconv4_stage3_L2 = mx.symbol.Convolution(name='Mconv4_stage3_L2', data=Mrelu3_stage3_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu4_stage3_L2 = mx.symbol.Activation(name='Mrelu4_stage3_L2', data=Mconv4_stage3_L2 , act_type='relu')
    Mconv5_stage3_L1 = mx.symbol.Convolution(name='Mconv5_stage3_L1', data=Mrelu4_stage3_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu5_stage3_L1 = mx.symbol.Activation(name='Mrelu5_stage3_L1', data=Mconv5_stage3_L1 , act_type='relu')
    Mconv5_stage3_L2 = mx.symbol.Convolution(name='Mconv5_stage3_L2', data=Mrelu4_stage3_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu5_stage3_L2 = mx.symbol.Activation(name='Mrelu5_stage3_L2', data=Mconv5_stage3_L2 , act_type='relu')
    Mconv6_stage3_L1 = mx.symbol.Convolution(name='Mconv6_stage3_L1', data=Mrelu5_stage3_L1 , num_filter=128, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mrelu6_stage3_L1 = mx.symbol.Activation(name='Mrelu6_stage3_L1', data=Mconv6_stage3_L1 , act_type='relu')
    Mconv6_stage3_L2 = mx.symbol.Convolution(name='Mconv6_stage3_L2', data=Mrelu5_stage3_L2 , num_filter=128, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mrelu6_stage3_L2 = mx.symbol.Activation(name='Mrelu6_stage3_L2', data=Mconv6_stage3_L2 , act_type='relu')
    Mconv7_stage3_L1 = mx.symbol.Convolution(name='Mconv7_stage3_L1', data=Mrelu6_stage3_L1 , num_filter=numoflinks*2, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mconv7_stage3_L2 = mx.symbol.Convolution(name='Mconv7_stage3_L2', data=Mrelu6_stage3_L2 , num_filter=numofparts, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    concat_stage4 = mx.symbol.Concat(name='concat_stage4', *[Mconv7_stage3_L1,Mconv7_stage3_L2,relu4_4_CPM] )
    Mconv1_stage4_L1 = mx.symbol.Convolution(name='Mconv1_stage4_L1', data=concat_stage4 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu1_stage4_L1 = mx.symbol.Activation(name='Mrelu1_stage4_L1', data=Mconv1_stage4_L1 , act_type='relu')
    Mconv1_stage4_L2 = mx.symbol.Convolution(name='Mconv1_stage4_L2', data=concat_stage4 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu1_stage4_L2 = mx.symbol.Activation(name='Mrelu1_stage4_L2', data=Mconv1_stage4_L2 , act_type='relu')
    Mconv2_stage4_L1 = mx.symbol.Convolution(name='Mconv2_stage4_L1', data=Mrelu1_stage4_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu2_stage4_L1 = mx.symbol.Activation(name='Mrelu2_stage4_L1', data=Mconv2_stage4_L1 , act_type='relu')
    Mconv2_stage4_L2 = mx.symbol.Convolution(name='Mconv2_stage4_L2', data=Mrelu1_stage4_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu2_stage4_L2 = mx.symbol.Activation(name='Mrelu2_stage4_L2', data=Mconv2_stage4_L2 , act_type='relu')
    Mconv3_stage4_L1 = mx.symbol.Convolution(name='Mconv3_stage4_L1', data=Mrelu2_stage4_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu3_stage4_L1 = mx.symbol.Activation(name='Mrelu3_stage4_L1', data=Mconv3_stage4_L1 , act_type='relu')
    Mconv3_stage4_L2 = mx.symbol.Convolution(name='Mconv3_stage4_L2', data=Mrelu2_stage4_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu3_stage4_L2 = mx.symbol.Activation(name='Mrelu3_stage4_L2', data=Mconv3_stage4_L2 , act_type='relu')
    Mconv4_stage4_L1 = mx.symbol.Convolution(name='Mconv4_stage4_L1', data=Mrelu3_stage4_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu4_stage4_L1 = mx.symbol.Activation(name='Mrelu4_stage4_L1', data=Mconv4_stage4_L1 , act_type='relu')
    Mconv4_stage4_L2 = mx.symbol.Convolution(name='Mconv4_stage4_L2', data=Mrelu3_stage4_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu4_stage4_L2 = mx.symbol.Activation(name='Mrelu4_stage4_L2', data=Mconv4_stage4_L2 , act_type='relu')
    Mconv5_stage4_L1 = mx.symbol.Convolution(name='Mconv5_stage4_L1', data=Mrelu4_stage4_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu5_stage4_L1 = mx.symbol.Activation(name='Mrelu5_stage4_L1', data=Mconv5_stage4_L1 , act_type='relu')
    Mconv5_stage4_L2 = mx.symbol.Convolution(name='Mconv5_stage4_L2', data=Mrelu4_stage4_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu5_stage4_L2 = mx.symbol.Activation(name='Mrelu5_stage4_L2', data=Mconv5_stage4_L2 , act_type='relu')
    Mconv6_stage4_L1 = mx.symbol.Convolution(name='Mconv6_stage4_L1', data=Mrelu5_stage4_L1 , num_filter=128, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mrelu6_stage4_L1 = mx.symbol.Activation(name='Mrelu6_stage4_L1', data=Mconv6_stage4_L1 , act_type='relu')
    Mconv6_stage4_L2 = mx.symbol.Convolution(name='Mconv6_stage4_L2', data=Mrelu5_stage4_L2 , num_filter=128, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mrelu6_stage4_L2 = mx.symbol.Activation(name='Mrelu6_stage4_L2', data=Mconv6_stage4_L2 , act_type='relu')
    Mconv7_stage4_L1 = mx.symbol.Convolution(name='Mconv7_stage4_L1', data=Mrelu6_stage4_L1 , num_filter=numoflinks*2, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mconv7_stage4_L2 = mx.symbol.Convolution(name='Mconv7_stage4_L2', data=Mrelu6_stage4_L2 , num_filter=numofparts, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    concat_stage5 = mx.symbol.Concat(name='concat_stage5', *[Mconv7_stage4_L1,Mconv7_stage4_L2,relu4_4_CPM] )
    Mconv1_stage5_L1 = mx.symbol.Convolution(name='Mconv1_stage5_L1', data=concat_stage5 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu1_stage5_L1 = mx.symbol.Activation(name='Mrelu1_stage5_L1', data=Mconv1_stage5_L1 , act_type='relu')
    Mconv1_stage5_L2 = mx.symbol.Convolution(name='Mconv1_stage5_L2', data=concat_stage5 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu1_stage5_L2 = mx.symbol.Activation(name='Mrelu1_stage5_L2', data=Mconv1_stage5_L2 , act_type='relu')
    Mconv2_stage5_L1 = mx.symbol.Convolution(name='Mconv2_stage5_L1', data=Mrelu1_stage5_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu2_stage5_L1 = mx.symbol.Activation(name='Mrelu2_stage5_L1', data=Mconv2_stage5_L1 , act_type='relu')
    Mconv2_stage5_L2 = mx.symbol.Convolution(name='Mconv2_stage5_L2', data=Mrelu1_stage5_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu2_stage5_L2 = mx.symbol.Activation(name='Mrelu2_stage5_L2', data=Mconv2_stage5_L2 , act_type='relu')
    Mconv3_stage5_L1 = mx.symbol.Convolution(name='Mconv3_stage5_L1', data=Mrelu2_stage5_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu3_stage5_L1 = mx.symbol.Activation(name='Mrelu3_stage5_L1', data=Mconv3_stage5_L1 , act_type='relu')
    Mconv3_stage5_L2 = mx.symbol.Convolution(name='Mconv3_stage5_L2', data=Mrelu2_stage5_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu3_stage5_L2 = mx.symbol.Activation(name='Mrelu3_stage5_L2', data=Mconv3_stage5_L2 , act_type='relu')
    Mconv4_stage5_L1 = mx.symbol.Convolution(name='Mconv4_stage5_L1', data=Mrelu3_stage5_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu4_stage5_L1 = mx.symbol.Activation(name='Mrelu4_stage5_L1', data=Mconv4_stage5_L1 , act_type='relu')
    Mconv4_stage5_L2 = mx.symbol.Convolution(name='Mconv4_stage5_L2', data=Mrelu3_stage5_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu4_stage5_L2 = mx.symbol.Activation(name='Mrelu4_stage5_L2', data=Mconv4_stage5_L2 , act_type='relu')
    Mconv5_stage5_L1 = mx.symbol.Convolution(name='Mconv5_stage5_L1', data=Mrelu4_stage5_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu5_stage5_L1 = mx.symbol.Activation(name='Mrelu5_stage5_L1', data=Mconv5_stage5_L1 , act_type='relu')
    Mconv5_stage5_L2 = mx.symbol.Convolution(name='Mconv5_stage5_L2', data=Mrelu4_stage5_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu5_stage5_L2 = mx.symbol.Activation(name='Mrelu5_stage5_L2', data=Mconv5_stage5_L2 , act_type='relu')
    Mconv6_stage5_L1 = mx.symbol.Convolution(name='Mconv6_stage5_L1', data=Mrelu5_stage5_L1 , num_filter=128, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mrelu6_stage5_L1 = mx.symbol.Activation(name='Mrelu6_stage5_L1', data=Mconv6_stage5_L1 , act_type='relu')
    Mconv6_stage5_L2 = mx.symbol.Convolution(name='Mconv6_stage5_L2', data=Mrelu5_stage5_L2 , num_filter=128, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mrelu6_stage5_L2 = mx.symbol.Activation(name='Mrelu6_stage5_L2', data=Mconv6_stage5_L2 , act_type='relu')
    Mconv7_stage5_L1 = mx.symbol.Convolution(name='Mconv7_stage5_L1', data=Mrelu6_stage5_L1 , num_filter=numoflinks*2, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mconv7_stage5_L2 = mx.symbol.Convolution(name='Mconv7_stage5_L2', data=Mrelu6_stage5_L2 , num_filter=numofparts, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    concat_stage6 = mx.symbol.Concat(name='concat_stage6', *[Mconv7_stage5_L1,Mconv7_stage5_L2,relu4_4_CPM] )
    Mconv1_stage6_L1 = mx.symbol.Convolution(name='Mconv1_stage6_L1', data=concat_stage6 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu1_stage6_L1 = mx.symbol.Activation(name='Mrelu1_stage6_L1', data=Mconv1_stage6_L1 , act_type='relu')
    Mconv1_stage6_L2 = mx.symbol.Convolution(name='Mconv1_stage6_L2', data=concat_stage6 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu1_stage6_L2 = mx.symbol.Activation(name='Mrelu1_stage6_L2', data=Mconv1_stage6_L2 , act_type='relu')
    Mconv2_stage6_L1 = mx.symbol.Convolution(name='Mconv2_stage6_L1', data=Mrelu1_stage6_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu2_stage6_L1 = mx.symbol.Activation(name='Mrelu2_stage6_L1', data=Mconv2_stage6_L1 , act_type='relu')
    Mconv2_stage6_L2 = mx.symbol.Convolution(name='Mconv2_stage6_L2', data=Mrelu1_stage6_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu2_stage6_L2 = mx.symbol.Activation(name='Mrelu2_stage6_L2', data=Mconv2_stage6_L2 , act_type='relu')
    Mconv3_stage6_L1 = mx.symbol.Convolution(name='Mconv3_stage6_L1', data=Mrelu2_stage6_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu3_stage6_L1 = mx.symbol.Activation(name='Mrelu3_stage6_L1', data=Mconv3_stage6_L1 , act_type='relu')
    Mconv3_stage6_L2 = mx.symbol.Convolution(name='Mconv3_stage6_L2', data=Mrelu2_stage6_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu3_stage6_L2 = mx.symbol.Activation(name='Mrelu3_stage6_L2', data=Mconv3_stage6_L2 , act_type='relu')
    Mconv4_stage6_L1 = mx.symbol.Convolution(name='Mconv4_stage6_L1', data=Mrelu3_stage6_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu4_stage6_L1 = mx.symbol.Activation(name='Mrelu4_stage6_L1', data=Mconv4_stage6_L1 , act_type='relu')
    Mconv4_stage6_L2 = mx.symbol.Convolution(name='Mconv4_stage6_L2', data=Mrelu3_stage6_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu4_stage6_L2 = mx.symbol.Activation(name='Mrelu4_stage6_L2', data=Mconv4_stage6_L2 , act_type='relu')
    Mconv5_stage6_L1 = mx.symbol.Convolution(name='Mconv5_stage6_L1', data=Mrelu4_stage6_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu5_stage6_L1 = mx.symbol.Activation(name='Mrelu5_stage6_L1', data=Mconv5_stage6_L1 , act_type='relu')
    Mconv5_stage6_L2 = mx.symbol.Convolution(name='Mconv5_stage6_L2', data=Mrelu4_stage6_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu5_stage6_L2 = mx.symbol.Activation(name='Mrelu5_stage6_L2', data=Mconv5_stage6_L2 , act_type='relu')
    Mconv6_stage6_L1 = mx.symbol.Convolution(name='Mconv6_stage6_L1', data=Mrelu5_stage6_L1 , num_filter=128, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mrelu6_stage6_L1 = mx.symbol.Activation(name='Mrelu6_stage6_L1', data=Mconv6_stage6_L1 , act_type='relu')
    Mconv6_stage6_L2 = mx.symbol.Convolution(name='Mconv6_stage6_L2', data=Mrelu5_stage6_L2 , num_filter=128, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mrelu6_stage6_L2 = mx.symbol.Activation(name='Mrelu6_stage6_L2', data=Mconv6_stage6_L2 , act_type='relu')
    Mconv7_stage6_L1 = mx.symbol.Convolution(name='Mconv7_stage6_L1', data=Mrelu6_stage6_L1 , num_filter=numoflinks*2, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mconv7_stage6_L2 = mx.symbol.Convolution(name='Mconv7_stage6_L2', data=Mrelu6_stage6_L2 , num_filter=numofparts, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)

    conv5_5_CPM_L1r = mx.symbol.Reshape(data=conv5_5_CPM_L1, shape=(-1,), name='conv5_5_CPM_L1r')
    partaffinityglabelr = mx.symbol.Reshape(data=partaffinityglabel, shape=(-1, ), name='partaffinityglabelr')
    stage1_loss_L1s = mx.symbol.square(conv5_5_CPM_L1r-partaffinityglabelr)
    vecweightw = mx.symbol.Reshape(data=vecweight, shape=(-1,), name='conv5_5_CPM_L1w')
    stage1_loss_L1w = stage1_loss_L1s*vecweightw
    stage1_loss_L1  = mx.symbol.MakeLoss(stage1_loss_L1w)
    
    conv5_5_CPM_L2r = mx.symbol.Reshape(data=conv5_5_CPM_L2, shape=(-1,), name='conv5_5_CPM_L2r')
    heatmaplabelr = mx.symbol.Reshape(data=heatmaplabel, shape=(-1, ), name='heatmaplabelr')
    stage1_loss_L2s = mx.symbol.square(conv5_5_CPM_L2r-heatmaplabelr)
    heatweightw = mx.symbol.Reshape(data=heatweight, shape=(-1,), name='conv5_5_CPM_L2w')
    stage1_loss_L2w = stage1_loss_L2s*heatweightw
    stage1_loss_L2  = mx.symbol.MakeLoss(stage1_loss_L2w)
        
    Mconv7_stage2_L1r = mx.symbol.Reshape(data=Mconv7_stage2_L1, shape=(-1,), name='Mconv7_stage2_L1')
    #partaffinityglabelr = mx.symbol.Reshape(data=partaffinityglabel, shape=(-1, ), name='partaffinityglabelr')
    stage2_loss_L1s = mx.symbol.square(Mconv7_stage2_L1r - partaffinityglabelr)
    #vecweightw = mx.symbol.Reshape(data=vecweight, shape=(-1,), name='Mconv7_stage2_L1r')
    stage2_loss_L1w = stage2_loss_L1s*vecweightw
    stage2_loss_L1  = mx.symbol.MakeLoss(stage2_loss_L1w)
    
    Mconv7_stage2_L2r = mx.symbol.Reshape(data=Mconv7_stage2_L2, shape=(-1,), name='Mconv7_stage2_L2')
    #heatmaplabelr = mx.symbol.Reshape(data=heatmaplabel, shape=(-1, ), name='heatmaplabelr')
    stage2_loss_L2s = mx.symbol.square(Mconv7_stage2_L2r-heatmaplabelr)
    #heatweightw = mx.symbol.Reshape(data=heatweight, shape=(-1,), name='conv5_5_CPM_L1r')
    stage2_loss_L2w = stage2_loss_L2s*heatweightw
    stage2_loss_L2  = mx.symbol.MakeLoss(stage2_loss_L2w)
    
    
    Mconv7_stage3_L1r = mx.symbol.Reshape(data=Mconv7_stage3_L1, shape=(-1,), name='Mconv7_stage3_L1')
    #partaffinityglabelr = mx.symbol.Reshape(data=partaffinityglabel, shape=(-1, ), name='partaffinityglabelr')
    stage3_loss_L1s = mx.symbol.square(Mconv7_stage3_L1r - partaffinityglabelr)
    #vecweightw = mx.symbol.Reshape(data=vecweight, shape=(-1,), name='Mconv7_stage2_L1r')
    stage3_loss_L1w = stage3_loss_L1s*vecweightw
    stage3_loss_L1  = mx.symbol.MakeLoss(stage3_loss_L1w)
    
    Mconv7_stage3_L2r = mx.symbol.Reshape(data=Mconv7_stage3_L2, shape=(-1,), name='Mconv7_stage3_L2')
    #heatmaplabelr = mx.symbol.Reshape(data=heatmaplabel, shape=(-1, ), name='heatmaplabelr')
    stage3_loss_L2s = mx.symbol.square(Mconv7_stage3_L2r-heatmaplabelr)
    #heatweightw = mx.symbol.Reshape(data=heatweight, shape=(-1,), name='conv5_5_CPM_L1r')
    stage3_loss_L2w = stage3_loss_L2s*heatweightw
    stage3_loss_L2  = mx.symbol.MakeLoss(stage3_loss_L2w)
    
    Mconv7_stage4_L1r = mx.symbol.Reshape(data=Mconv7_stage4_L1, shape=(-1,), name='Mconv7_stage4_L1')
    #partaffinityglabelr = mx.symbol.Reshape(data=partaffinityglabel, shape=(-1, ), name='partaffinityglabelr')
    stage4_loss_L1s = mx.symbol.square(Mconv7_stage4_L1r - partaffinityglabelr)
    #vecweightw = mx.symbol.Reshape(data=vecweight, shape=(-1,), name='Mconv7_stage2_L1r')
    stage4_loss_L1w = stage4_loss_L1s*vecweightw
    stage4_loss_L1  = mx.symbol.MakeLoss(stage4_loss_L1w)
    
    Mconv7_stage4_L2r = mx.symbol.Reshape(data=Mconv7_stage4_L2, shape=(-1,), name='Mconv7_stage4_L2')
    #heatmaplabelr = mx.symbol.Reshape(data=heatmaplabel, shape=(-1, ), name='heatmaplabelr')
    stage4_loss_L2s = mx.symbol.square(Mconv7_stage4_L2r-heatmaplabelr)
    #heatweightw = mx.symbol.Reshape(data=heatweight, shape=(-1,), name='conv5_5_CPM_L1r')
    stage4_loss_L2w = stage4_loss_L2s*heatweightw
    stage4_loss_L2  = mx.symbol.MakeLoss(stage4_loss_L2w)
    
    Mconv7_stage5_L1r = mx.symbol.Reshape(data=Mconv7_stage5_L1, shape=(-1,), name='Mconv7_stage5_L1')
    #partaffinityglabelr = mx.symbol.Reshape(data=partaffinityglabel, shape=(-1, ), name='partaffinityglabelr')
    stage5_loss_L1s = mx.symbol.square(Mconv7_stage5_L1r - partaffinityglabelr)
    #vecweightw = mx.symbol.Reshape(data=vecweight, shape=(-1,), name='Mconv7_stage2_L1r')
    stage5_loss_L1w = stage5_loss_L1s*vecweightw
    stage5_loss_L1  = mx.symbol.MakeLoss(stage5_loss_L1w)
    
    Mconv7_stage5_L2r = mx.symbol.Reshape(data=Mconv7_stage5_L2, shape=(-1,), name='Mconv7_stage5_L2')
    #heatmaplabelr = mx.symbol.Reshape(data=heatmaplabel, shape=(-1, ), name='heatmaplabelr')
    stage5_loss_L2s = mx.symbol.square(Mconv7_stage5_L2r-heatmaplabelr)
    #heatweightw = mx.symbol.Reshape(data=heatweight, shape=(-1,), name='conv5_5_CPM_L1r')
    stage5_loss_L2w = stage5_loss_L2s*heatweightw
    stage5_loss_L2  = mx.symbol.MakeLoss(stage5_loss_L2w)
    
    
    Mconv7_stage6_L1r = mx.symbol.Reshape(data=Mconv7_stage6_L1, shape=(-1,), name='Mconv7_stage3_L1')
    #partaffinityglabelr = mx.symbol.Reshape(data=partaffinityglabel, shape=(-1, ), name='partaffinityglabelr')
    stage6_loss_L1s = mx.symbol.square(Mconv7_stage6_L1r - partaffinityglabelr)
    #vecweightw = mx.symbol.Reshape(data=vecweight, shape=(-1,), name='Mconv7_stage2_L1r')
    stage6_loss_L1w = stage6_loss_L1s*vecweightw
    stage6_loss_L1  = mx.symbol.MakeLoss(stage6_loss_L1w)
    
    Mconv7_stage6_L2r = mx.symbol.Reshape(data=Mconv7_stage6_L2, shape=(-1,), name='Mconv7_stage3_L2')
    #heatmaplabelr = mx.symbol.Reshape(data=heatmaplabel, shape=(-1, ), name='heatmaplabelr')
    stage6_loss_L2s = mx.symbol.square(Mconv7_stage6_L2r-heatmaplabelr)
    #heatweightw = mx.symbol.Reshape(data=heatweight, shape=(-1,), name='conv5_5_CPM_L1r')
    stage6_loss_L2w = stage6_loss_L2s*heatweightw
    stage6_loss_L2  = mx.symbol.MakeLoss(stage6_loss_L2w)
    
    group = mx.symbol.Group([stage1_loss_L1, stage1_loss_L2,
                             stage2_loss_L1, stage2_loss_L2,
                             stage3_loss_L1, stage3_loss_L2,
                             stage4_loss_L1, stage4_loss_L2,
                             stage5_loss_L1, stage5_loss_L2,
                             stage6_loss_L1, stage6_loss_L2])
    return group

def CPMModel_test():
   
    data = mx.symbol.Variable(name='data')
    ## heat map of human parts
    heatmaplabel = mx.sym.Variable("heatmaplabel")
    ## part affinity graph
    partaffinityglabel = mx.sym.Variable('partaffinityglabel')

    heatweight = mx.sym.Variable('heatweight')
    
    vecweight = mx.sym.Variable('vecweight')
    
    conv1_1 = mx.symbol.Convolution(name='conv1_1', data=data , num_filter=64, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu1_1 = mx.symbol.Activation(name='relu1_1', data=conv1_1 , act_type='relu')
    conv1_2 = mx.symbol.Convolution(name='conv1_2', data=relu1_1 , num_filter=64, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu1_2 = mx.symbol.Activation(name='relu1_2', data=conv1_2 , act_type='relu')
    pool1_stage1 = mx.symbol.Pooling(name='pool1_stage1', data=relu1_2 , pooling_convention='full', pad=(0,0), kernel=(2,2), stride=(2,2), pool_type='max')
    conv2_1 = mx.symbol.Convolution(name='conv2_1', data=pool1_stage1 , num_filter=128, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu2_1 = mx.symbol.Activation(name='relu2_1', data=conv2_1 , act_type='relu')
    conv2_2 = mx.symbol.Convolution(name='conv2_2', data=relu2_1 , num_filter=128, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu2_2 = mx.symbol.Activation(name='relu2_2', data=conv2_2 , act_type='relu')
    pool2_stage1 = mx.symbol.Pooling(name='pool2_stage1', data=relu2_2 , pooling_convention='full', pad=(0,0), kernel=(2,2), stride=(2,2), pool_type='max')
    conv3_1 = mx.symbol.Convolution(name='conv3_1', data=pool2_stage1 , num_filter=256, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu3_1 = mx.symbol.Activation(name='relu3_1', data=conv3_1 , act_type='relu')
    conv3_2 = mx.symbol.Convolution(name='conv3_2', data=relu3_1 , num_filter=256, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu3_2 = mx.symbol.Activation(name='relu3_2', data=conv3_2 , act_type='relu')
    conv3_3 = mx.symbol.Convolution(name='conv3_3', data=relu3_2 , num_filter=256, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu3_3 = mx.symbol.Activation(name='relu3_3', data=conv3_3 , act_type='relu')
    conv3_4 = mx.symbol.Convolution(name='conv3_4', data=relu3_3 , num_filter=256, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu3_4 = mx.symbol.Activation(name='relu3_4', data=conv3_4 , act_type='relu')
    pool3_stage1 = mx.symbol.Pooling(name='pool3_stage1', data=relu3_4 , pooling_convention='full', pad=(0,0), kernel=(2,2), stride=(2,2), pool_type='max')
    conv4_1 = mx.symbol.Convolution(name='conv4_1', data=pool3_stage1 , num_filter=512, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu4_1 = mx.symbol.Activation(name='relu4_1', data=conv4_1 , act_type='relu')
    conv4_2 = mx.symbol.Convolution(name='conv4_2', data=relu4_1 , num_filter=512, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu4_2 = mx.symbol.Activation(name='relu4_2', data=conv4_2 , act_type='relu')
    conv4_3_CPM = mx.symbol.Convolution(name='conv4_3_CPM', data=relu4_2 , num_filter=256, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu4_3_CPM = mx.symbol.Activation(name='relu4_3_CPM', data=conv4_3_CPM , act_type='relu')
    conv4_4_CPM = mx.symbol.Convolution(name='conv4_4_CPM', data=relu4_3_CPM , num_filter=128, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu4_4_CPM = mx.symbol.Activation(name='relu4_4_CPM', data=conv4_4_CPM , act_type='relu')
    conv5_1_CPM_L1 = mx.symbol.Convolution(name='conv5_1_CPM_L1', data=relu4_4_CPM , num_filter=128, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu5_1_CPM_L1 = mx.symbol.Activation(name='relu5_1_CPM_L1', data=conv5_1_CPM_L1 , act_type='relu')
    conv5_1_CPM_L2 = mx.symbol.Convolution(name='conv5_1_CPM_L2', data=relu4_4_CPM , num_filter=128, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu5_1_CPM_L2 = mx.symbol.Activation(name='relu5_1_CPM_L2', data=conv5_1_CPM_L2 , act_type='relu')
    conv5_2_CPM_L1 = mx.symbol.Convolution(name='conv5_2_CPM_L1', data=relu5_1_CPM_L1 , num_filter=128, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu5_2_CPM_L1 = mx.symbol.Activation(name='relu5_2_CPM_L1', data=conv5_2_CPM_L1 , act_type='relu')
    conv5_2_CPM_L2 = mx.symbol.Convolution(name='conv5_2_CPM_L2', data=relu5_1_CPM_L2 , num_filter=128, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu5_2_CPM_L2 = mx.symbol.Activation(name='relu5_2_CPM_L2', data=conv5_2_CPM_L2 , act_type='relu')
    conv5_3_CPM_L1 = mx.symbol.Convolution(name='conv5_3_CPM_L1', data=relu5_2_CPM_L1 , num_filter=128, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu5_3_CPM_L1 = mx.symbol.Activation(name='relu5_3_CPM_L1', data=conv5_3_CPM_L1 , act_type='relu')
    conv5_3_CPM_L2 = mx.symbol.Convolution(name='conv5_3_CPM_L2', data=relu5_2_CPM_L2 , num_filter=128, pad=(1,1), kernel=(3,3), stride=(1,1), no_bias=False)
    relu5_3_CPM_L2 = mx.symbol.Activation(name='relu5_3_CPM_L2', data=conv5_3_CPM_L2 , act_type='relu')
    conv5_4_CPM_L1 = mx.symbol.Convolution(name='conv5_4_CPM_L1', data=relu5_3_CPM_L1 , num_filter=512, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    relu5_4_CPM_L1 = mx.symbol.Activation(name='relu5_4_CPM_L1', data=conv5_4_CPM_L1 , act_type='relu')
    conv5_4_CPM_L2 = mx.symbol.Convolution(name='conv5_4_CPM_L2', data=relu5_3_CPM_L2 , num_filter=512, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    relu5_4_CPM_L2 = mx.symbol.Activation(name='relu5_4_CPM_L2', data=conv5_4_CPM_L2 , act_type='relu')
    conv5_5_CPM_L1 = mx.symbol.Convolution(name='conv5_5_CPM_L1', data=relu5_4_CPM_L1 , num_filter=numoflinks*2, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    conv5_5_CPM_L2 = mx.symbol.Convolution(name='conv5_5_CPM_L2', data=relu5_4_CPM_L2 , num_filter=numofparts, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    concat_stage2 = mx.symbol.Concat(name='concat_stage2', *[conv5_5_CPM_L1,conv5_5_CPM_L2,relu4_4_CPM] )
    Mconv1_stage2_L1 = mx.symbol.Convolution(name='Mconv1_stage2_L1', data=concat_stage2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu1_stage2_L1 = mx.symbol.Activation(name='Mrelu1_stage2_L1', data=Mconv1_stage2_L1 , act_type='relu')
    Mconv1_stage2_L2 = mx.symbol.Convolution(name='Mconv1_stage2_L2', data=concat_stage2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu1_stage2_L2 = mx.symbol.Activation(name='Mrelu1_stage2_L2', data=Mconv1_stage2_L2 , act_type='relu')
    Mconv2_stage2_L1 = mx.symbol.Convolution(name='Mconv2_stage2_L1', data=Mrelu1_stage2_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu2_stage2_L1 = mx.symbol.Activation(name='Mrelu2_stage2_L1', data=Mconv2_stage2_L1 , act_type='relu')
    Mconv2_stage2_L2 = mx.symbol.Convolution(name='Mconv2_stage2_L2', data=Mrelu1_stage2_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu2_stage2_L2 = mx.symbol.Activation(name='Mrelu2_stage2_L2', data=Mconv2_stage2_L2 , act_type='relu')
    Mconv3_stage2_L1 = mx.symbol.Convolution(name='Mconv3_stage2_L1', data=Mrelu2_stage2_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu3_stage2_L1 = mx.symbol.Activation(name='Mrelu3_stage2_L1', data=Mconv3_stage2_L1 , act_type='relu')
    Mconv3_stage2_L2 = mx.symbol.Convolution(name='Mconv3_stage2_L2', data=Mrelu2_stage2_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu3_stage2_L2 = mx.symbol.Activation(name='Mrelu3_stage2_L2', data=Mconv3_stage2_L2 , act_type='relu')
    Mconv4_stage2_L1 = mx.symbol.Convolution(name='Mconv4_stage2_L1', data=Mrelu3_stage2_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu4_stage2_L1 = mx.symbol.Activation(name='Mrelu4_stage2_L1', data=Mconv4_stage2_L1 , act_type='relu')
    Mconv4_stage2_L2 = mx.symbol.Convolution(name='Mconv4_stage2_L2', data=Mrelu3_stage2_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu4_stage2_L2 = mx.symbol.Activation(name='Mrelu4_stage2_L2', data=Mconv4_stage2_L2 , act_type='relu')
    Mconv5_stage2_L1 = mx.symbol.Convolution(name='Mconv5_stage2_L1', data=Mrelu4_stage2_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu5_stage2_L1 = mx.symbol.Activation(name='Mrelu5_stage2_L1', data=Mconv5_stage2_L1 , act_type='relu')
    Mconv5_stage2_L2 = mx.symbol.Convolution(name='Mconv5_stage2_L2', data=Mrelu4_stage2_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu5_stage2_L2 = mx.symbol.Activation(name='Mrelu5_stage2_L2', data=Mconv5_stage2_L2 , act_type='relu')
    Mconv6_stage2_L1 = mx.symbol.Convolution(name='Mconv6_stage2_L1', data=Mrelu5_stage2_L1 , num_filter=128, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mrelu6_stage2_L1 = mx.symbol.Activation(name='Mrelu6_stage2_L1', data=Mconv6_stage2_L1 , act_type='relu')
    Mconv6_stage2_L2 = mx.symbol.Convolution(name='Mconv6_stage2_L2', data=Mrelu5_stage2_L2 , num_filter=128, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mrelu6_stage2_L2 = mx.symbol.Activation(name='Mrelu6_stage2_L2', data=Mconv6_stage2_L2 , act_type='relu')
    Mconv7_stage2_L1 = mx.symbol.Convolution(name='Mconv7_stage2_L1', data=Mrelu6_stage2_L1 , num_filter=numoflinks*2, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mconv7_stage2_L2 = mx.symbol.Convolution(name='Mconv7_stage2_L2', data=Mrelu6_stage2_L2 , num_filter=numofparts, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    concat_stage3 = mx.symbol.Concat(name='concat_stage3', *[Mconv7_stage2_L1,Mconv7_stage2_L2,relu4_4_CPM] )
    Mconv1_stage3_L1 = mx.symbol.Convolution(name='Mconv1_stage3_L1', data=concat_stage3 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu1_stage3_L1 = mx.symbol.Activation(name='Mrelu1_stage3_L1', data=Mconv1_stage3_L1 , act_type='relu')
    Mconv1_stage3_L2 = mx.symbol.Convolution(name='Mconv1_stage3_L2', data=concat_stage3 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu1_stage3_L2 = mx.symbol.Activation(name='Mrelu1_stage3_L2', data=Mconv1_stage3_L2 , act_type='relu')
    Mconv2_stage3_L1 = mx.symbol.Convolution(name='Mconv2_stage3_L1', data=Mrelu1_stage3_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu2_stage3_L1 = mx.symbol.Activation(name='Mrelu2_stage3_L1', data=Mconv2_stage3_L1 , act_type='relu')
    Mconv2_stage3_L2 = mx.symbol.Convolution(name='Mconv2_stage3_L2', data=Mrelu1_stage3_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu2_stage3_L2 = mx.symbol.Activation(name='Mrelu2_stage3_L2', data=Mconv2_stage3_L2 , act_type='relu')
    Mconv3_stage3_L1 = mx.symbol.Convolution(name='Mconv3_stage3_L1', data=Mrelu2_stage3_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu3_stage3_L1 = mx.symbol.Activation(name='Mrelu3_stage3_L1', data=Mconv3_stage3_L1 , act_type='relu')
    Mconv3_stage3_L2 = mx.symbol.Convolution(name='Mconv3_stage3_L2', data=Mrelu2_stage3_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu3_stage3_L2 = mx.symbol.Activation(name='Mrelu3_stage3_L2', data=Mconv3_stage3_L2 , act_type='relu')
    Mconv4_stage3_L1 = mx.symbol.Convolution(name='Mconv4_stage3_L1', data=Mrelu3_stage3_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu4_stage3_L1 = mx.symbol.Activation(name='Mrelu4_stage3_L1', data=Mconv4_stage3_L1 , act_type='relu')
    Mconv4_stage3_L2 = mx.symbol.Convolution(name='Mconv4_stage3_L2', data=Mrelu3_stage3_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu4_stage3_L2 = mx.symbol.Activation(name='Mrelu4_stage3_L2', data=Mconv4_stage3_L2 , act_type='relu')
    Mconv5_stage3_L1 = mx.symbol.Convolution(name='Mconv5_stage3_L1', data=Mrelu4_stage3_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu5_stage3_L1 = mx.symbol.Activation(name='Mrelu5_stage3_L1', data=Mconv5_stage3_L1 , act_type='relu')
    Mconv5_stage3_L2 = mx.symbol.Convolution(name='Mconv5_stage3_L2', data=Mrelu4_stage3_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu5_stage3_L2 = mx.symbol.Activation(name='Mrelu5_stage3_L2', data=Mconv5_stage3_L2 , act_type='relu')
    Mconv6_stage3_L1 = mx.symbol.Convolution(name='Mconv6_stage3_L1', data=Mrelu5_stage3_L1 , num_filter=128, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mrelu6_stage3_L1 = mx.symbol.Activation(name='Mrelu6_stage3_L1', data=Mconv6_stage3_L1 , act_type='relu')
    Mconv6_stage3_L2 = mx.symbol.Convolution(name='Mconv6_stage3_L2', data=Mrelu5_stage3_L2 , num_filter=128, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mrelu6_stage3_L2 = mx.symbol.Activation(name='Mrelu6_stage3_L2', data=Mconv6_stage3_L2 , act_type='relu')
    Mconv7_stage3_L1 = mx.symbol.Convolution(name='Mconv7_stage3_L1', data=Mrelu6_stage3_L1 , num_filter=numoflinks*2, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mconv7_stage3_L2 = mx.symbol.Convolution(name='Mconv7_stage3_L2', data=Mrelu6_stage3_L2 , num_filter=numofparts, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    concat_stage4 = mx.symbol.Concat(name='concat_stage4', *[Mconv7_stage3_L1,Mconv7_stage3_L2,relu4_4_CPM] )
    Mconv1_stage4_L1 = mx.symbol.Convolution(name='Mconv1_stage4_L1', data=concat_stage4 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu1_stage4_L1 = mx.symbol.Activation(name='Mrelu1_stage4_L1', data=Mconv1_stage4_L1 , act_type='relu')
    Mconv1_stage4_L2 = mx.symbol.Convolution(name='Mconv1_stage4_L2', data=concat_stage4 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu1_stage4_L2 = mx.symbol.Activation(name='Mrelu1_stage4_L2', data=Mconv1_stage4_L2 , act_type='relu')
    Mconv2_stage4_L1 = mx.symbol.Convolution(name='Mconv2_stage4_L1', data=Mrelu1_stage4_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu2_stage4_L1 = mx.symbol.Activation(name='Mrelu2_stage4_L1', data=Mconv2_stage4_L1 , act_type='relu')
    Mconv2_stage4_L2 = mx.symbol.Convolution(name='Mconv2_stage4_L2', data=Mrelu1_stage4_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu2_stage4_L2 = mx.symbol.Activation(name='Mrelu2_stage4_L2', data=Mconv2_stage4_L2 , act_type='relu')
    Mconv3_stage4_L1 = mx.symbol.Convolution(name='Mconv3_stage4_L1', data=Mrelu2_stage4_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu3_stage4_L1 = mx.symbol.Activation(name='Mrelu3_stage4_L1', data=Mconv3_stage4_L1 , act_type='relu')
    Mconv3_stage4_L2 = mx.symbol.Convolution(name='Mconv3_stage4_L2', data=Mrelu2_stage4_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu3_stage4_L2 = mx.symbol.Activation(name='Mrelu3_stage4_L2', data=Mconv3_stage4_L2 , act_type='relu')
    Mconv4_stage4_L1 = mx.symbol.Convolution(name='Mconv4_stage4_L1', data=Mrelu3_stage4_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu4_stage4_L1 = mx.symbol.Activation(name='Mrelu4_stage4_L1', data=Mconv4_stage4_L1 , act_type='relu')
    Mconv4_stage4_L2 = mx.symbol.Convolution(name='Mconv4_stage4_L2', data=Mrelu3_stage4_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu4_stage4_L2 = mx.symbol.Activation(name='Mrelu4_stage4_L2', data=Mconv4_stage4_L2 , act_type='relu')
    Mconv5_stage4_L1 = mx.symbol.Convolution(name='Mconv5_stage4_L1', data=Mrelu4_stage4_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu5_stage4_L1 = mx.symbol.Activation(name='Mrelu5_stage4_L1', data=Mconv5_stage4_L1 , act_type='relu')
    Mconv5_stage4_L2 = mx.symbol.Convolution(name='Mconv5_stage4_L2', data=Mrelu4_stage4_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu5_stage4_L2 = mx.symbol.Activation(name='Mrelu5_stage4_L2', data=Mconv5_stage4_L2 , act_type='relu')
    Mconv6_stage4_L1 = mx.symbol.Convolution(name='Mconv6_stage4_L1', data=Mrelu5_stage4_L1 , num_filter=128, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mrelu6_stage4_L1 = mx.symbol.Activation(name='Mrelu6_stage4_L1', data=Mconv6_stage4_L1 , act_type='relu')
    Mconv6_stage4_L2 = mx.symbol.Convolution(name='Mconv6_stage4_L2', data=Mrelu5_stage4_L2 , num_filter=128, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mrelu6_stage4_L2 = mx.symbol.Activation(name='Mrelu6_stage4_L2', data=Mconv6_stage4_L2 , act_type='relu')
    Mconv7_stage4_L1 = mx.symbol.Convolution(name='Mconv7_stage4_L1', data=Mrelu6_stage4_L1 , num_filter=numoflinks*2, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mconv7_stage4_L2 = mx.symbol.Convolution(name='Mconv7_stage4_L2', data=Mrelu6_stage4_L2 , num_filter=numofparts, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    concat_stage5 = mx.symbol.Concat(name='concat_stage5', *[Mconv7_stage4_L1,Mconv7_stage4_L2,relu4_4_CPM] )
    Mconv1_stage5_L1 = mx.symbol.Convolution(name='Mconv1_stage5_L1', data=concat_stage5 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu1_stage5_L1 = mx.symbol.Activation(name='Mrelu1_stage5_L1', data=Mconv1_stage5_L1 , act_type='relu')
    Mconv1_stage5_L2 = mx.symbol.Convolution(name='Mconv1_stage5_L2', data=concat_stage5 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu1_stage5_L2 = mx.symbol.Activation(name='Mrelu1_stage5_L2', data=Mconv1_stage5_L2 , act_type='relu')
    Mconv2_stage5_L1 = mx.symbol.Convolution(name='Mconv2_stage5_L1', data=Mrelu1_stage5_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu2_stage5_L1 = mx.symbol.Activation(name='Mrelu2_stage5_L1', data=Mconv2_stage5_L1 , act_type='relu')
    Mconv2_stage5_L2 = mx.symbol.Convolution(name='Mconv2_stage5_L2', data=Mrelu1_stage5_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu2_stage5_L2 = mx.symbol.Activation(name='Mrelu2_stage5_L2', data=Mconv2_stage5_L2 , act_type='relu')
    Mconv3_stage5_L1 = mx.symbol.Convolution(name='Mconv3_stage5_L1', data=Mrelu2_stage5_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu3_stage5_L1 = mx.symbol.Activation(name='Mrelu3_stage5_L1', data=Mconv3_stage5_L1 , act_type='relu')
    Mconv3_stage5_L2 = mx.symbol.Convolution(name='Mconv3_stage5_L2', data=Mrelu2_stage5_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu3_stage5_L2 = mx.symbol.Activation(name='Mrelu3_stage5_L2', data=Mconv3_stage5_L2 , act_type='relu')
    Mconv4_stage5_L1 = mx.symbol.Convolution(name='Mconv4_stage5_L1', data=Mrelu3_stage5_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu4_stage5_L1 = mx.symbol.Activation(name='Mrelu4_stage5_L1', data=Mconv4_stage5_L1 , act_type='relu')
    Mconv4_stage5_L2 = mx.symbol.Convolution(name='Mconv4_stage5_L2', data=Mrelu3_stage5_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu4_stage5_L2 = mx.symbol.Activation(name='Mrelu4_stage5_L2', data=Mconv4_stage5_L2 , act_type='relu')
    Mconv5_stage5_L1 = mx.symbol.Convolution(name='Mconv5_stage5_L1', data=Mrelu4_stage5_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu5_stage5_L1 = mx.symbol.Activation(name='Mrelu5_stage5_L1', data=Mconv5_stage5_L1 , act_type='relu')
    Mconv5_stage5_L2 = mx.symbol.Convolution(name='Mconv5_stage5_L2', data=Mrelu4_stage5_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu5_stage5_L2 = mx.symbol.Activation(name='Mrelu5_stage5_L2', data=Mconv5_stage5_L2 , act_type='relu')
    Mconv6_stage5_L1 = mx.symbol.Convolution(name='Mconv6_stage5_L1', data=Mrelu5_stage5_L1 , num_filter=128, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mrelu6_stage5_L1 = mx.symbol.Activation(name='Mrelu6_stage5_L1', data=Mconv6_stage5_L1 , act_type='relu')
    Mconv6_stage5_L2 = mx.symbol.Convolution(name='Mconv6_stage5_L2', data=Mrelu5_stage5_L2 , num_filter=128, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mrelu6_stage5_L2 = mx.symbol.Activation(name='Mrelu6_stage5_L2', data=Mconv6_stage5_L2 , act_type='relu')
    Mconv7_stage5_L1 = mx.symbol.Convolution(name='Mconv7_stage5_L1', data=Mrelu6_stage5_L1 , num_filter=numoflinks*2, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mconv7_stage5_L2 = mx.symbol.Convolution(name='Mconv7_stage5_L2', data=Mrelu6_stage5_L2 , num_filter=numofparts, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    concat_stage6 = mx.symbol.Concat(name='concat_stage6', *[Mconv7_stage5_L1,Mconv7_stage5_L2,relu4_4_CPM] )
    Mconv1_stage6_L1 = mx.symbol.Convolution(name='Mconv1_stage6_L1', data=concat_stage6 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu1_stage6_L1 = mx.symbol.Activation(name='Mrelu1_stage6_L1', data=Mconv1_stage6_L1 , act_type='relu')
    Mconv1_stage6_L2 = mx.symbol.Convolution(name='Mconv1_stage6_L2', data=concat_stage6 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu1_stage6_L2 = mx.symbol.Activation(name='Mrelu1_stage6_L2', data=Mconv1_stage6_L2 , act_type='relu')
    Mconv2_stage6_L1 = mx.symbol.Convolution(name='Mconv2_stage6_L1', data=Mrelu1_stage6_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu2_stage6_L1 = mx.symbol.Activation(name='Mrelu2_stage6_L1', data=Mconv2_stage6_L1 , act_type='relu')
    Mconv2_stage6_L2 = mx.symbol.Convolution(name='Mconv2_stage6_L2', data=Mrelu1_stage6_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu2_stage6_L2 = mx.symbol.Activation(name='Mrelu2_stage6_L2', data=Mconv2_stage6_L2 , act_type='relu')
    Mconv3_stage6_L1 = mx.symbol.Convolution(name='Mconv3_stage6_L1', data=Mrelu2_stage6_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu3_stage6_L1 = mx.symbol.Activation(name='Mrelu3_stage6_L1', data=Mconv3_stage6_L1 , act_type='relu')
    Mconv3_stage6_L2 = mx.symbol.Convolution(name='Mconv3_stage6_L2', data=Mrelu2_stage6_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu3_stage6_L2 = mx.symbol.Activation(name='Mrelu3_stage6_L2', data=Mconv3_stage6_L2 , act_type='relu')
    Mconv4_stage6_L1 = mx.symbol.Convolution(name='Mconv4_stage6_L1', data=Mrelu3_stage6_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu4_stage6_L1 = mx.symbol.Activation(name='Mrelu4_stage6_L1', data=Mconv4_stage6_L1 , act_type='relu')
    Mconv4_stage6_L2 = mx.symbol.Convolution(name='Mconv4_stage6_L2', data=Mrelu3_stage6_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu4_stage6_L2 = mx.symbol.Activation(name='Mrelu4_stage6_L2', data=Mconv4_stage6_L2 , act_type='relu')
    Mconv5_stage6_L1 = mx.symbol.Convolution(name='Mconv5_stage6_L1', data=Mrelu4_stage6_L1 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu5_stage6_L1 = mx.symbol.Activation(name='Mrelu5_stage6_L1', data=Mconv5_stage6_L1 , act_type='relu')
    Mconv5_stage6_L2 = mx.symbol.Convolution(name='Mconv5_stage6_L2', data=Mrelu4_stage6_L2 , num_filter=128, pad=(3,3), kernel=(7,7), stride=(1,1), no_bias=False)
    Mrelu5_stage6_L2 = mx.symbol.Activation(name='Mrelu5_stage6_L2', data=Mconv5_stage6_L2 , act_type='relu')
    Mconv6_stage6_L1 = mx.symbol.Convolution(name='Mconv6_stage6_L1', data=Mrelu5_stage6_L1 , num_filter=128, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mrelu6_stage6_L1 = mx.symbol.Activation(name='Mrelu6_stage6_L1', data=Mconv6_stage6_L1 , act_type='relu')
    Mconv6_stage6_L2 = mx.symbol.Convolution(name='Mconv6_stage6_L2', data=Mrelu5_stage6_L2 , num_filter=128, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mrelu6_stage6_L2 = mx.symbol.Activation(name='Mrelu6_stage6_L2', data=Mconv6_stage6_L2 , act_type='relu')
    Mconv7_stage6_L1 = mx.symbol.Convolution(name='Mconv7_stage6_L1', data=Mrelu6_stage6_L1 , num_filter=numoflinks*2, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)
    Mconv7_stage6_L2 = mx.symbol.Convolution(name='Mconv7_stage6_L2', data=Mrelu6_stage6_L2 , num_filter=numofparts, pad=(0,0), kernel=(1,1), stride=(1,1), no_bias=False)

    
    group = mx.symbol.Group([conv5_5_CPM_L1, conv5_5_CPM_L2,
                             Mconv7_stage2_L1, Mconv7_stage2_L2,
                             Mconv7_stage3_L1, Mconv7_stage3_L2,
                             Mconv7_stage4_L1, Mconv7_stage4_L2,
                             Mconv7_stage5_L1, Mconv7_stage5_L2,
                             Mconv7_stage6_L1, Mconv7_stage6_L2])
    return group

class DataBatchweight(object):
    def __init__(self, data, heatmaplabel, partaffinityglabel, heatweight, vecweight, pad=0):
        self.data = [data]
        self.label = [heatmaplabel, partaffinityglabel, heatweight, vecweight]
        self.pad = pad