Community
cancel
Showing results for 
Search instead for 
Did you mean: 
Highlighted
Community Manager
240 Views

[Error 5] Toolkit Error: Stage Details Not Supported: Equal

Hi,

 

I want to run a custom model implemented in Tensorflow on Cifar 10 dataset. However when I run this command : mvNCCompile -s 12 '100.meta' -o cifar_10_model -in 'data/Input' -on 'output1/output1' it fails I get this error : [Error 5] Toolkit Error: Stage Details Not Supported: Equal

 

My model is :

 

`import tensorflow as tf

 

def model():

 

_IMAGE_SIZE = 32

 

_IMAGE_CHANNELS = 3

 

_NUM_CLASSES = 10

 

_RESHAPE_SIZE = 4_4_128

 

with tf.name_scope('data'): x = tf.placeholder(tf.float32, shape=[None, _IMAGE_SIZE * _IMAGE_SIZE * _IMAGE_CHANNELS], name='Input') y = tf.placeholder(tf.float32, shape=[None, _NUM_CLASSES], name='Output') x_image = tf.reshape(x, [-1, _IMAGE_SIZE, _IMAGE_SIZE, _IMAGE_CHANNELS], name='images') def variable_with_weight_decay(name, shape, stddev, wd): dtype = tf.float32 var = variable_on_cpu( name, shape, tf.truncated_normal_initializer(stddev=stddev, dtype=dtype)) if wd is not None: weight_decay = tf.multiply(tf.nn.l2_loss(var), wd, name='weight_loss') tf.add_to_collection('losses', weight_decay) return var def variable_on_cpu(name, shape, initializer): with tf.device('/cpu:0'): dtype = tf.float32 var = tf.get_variable(name, shape, initializer=initializer, dtype=dtype) return var with tf.variable_scope('conv1') as scope: kernel = variable_with_weight_decay('weights', shape=[5, 5, 3, 64], stddev=5e-2, wd=0.0) conv = tf.nn.conv2d(x_image, kernel, [1, 1, 1, 1], padding='SAME') biases = variable_on_cpu('biases', [64], tf.constant_initializer(0.0)) pre_activation = tf.nn.bias_add(conv, biases) conv1 = tf.nn.relu(pre_activation, name=scope.name) #tf.summary.histogram('Convolution_layers/conv1', conv1) tf.summary.scalar('Convolution_layers/conv1', tf.nn.zero_fraction(conv1)) norm1 = tf.nn.lrn(conv1, 4, bias=1.0, alpha=0.001 / 9.0, beta=0.75, name='norm1') pool1 = tf.nn.max_pool(norm1, ksize=[1, 3, 3, 1], strides=[1, 2, 2, 1], padding='SAME', name='pool1') with tf.variable_scope('conv2') as scope: kernel = variable_with_weight_decay('weights', shape=[5, 5, 64, 64], stddev=5e-2, wd=0.0) conv = tf.nn.conv2d(pool1, kernel, [1, 1, 1, 1], padding='SAME') biases = variable_on_cpu('biases', [64], tf.constant_initializer(0.1)) pre_activation = tf.nn.bias_add(conv, biases) conv2 = tf.nn.relu(pre_activation, name=scope.name) #tf.summary.histogram('Convolution_layers/conv2', conv2) tf.summary.scalar('Convolution_layers/conv2', tf.nn.zero_fraction(conv2)) norm2 = tf.nn.lrn(conv2, 4, bias=1.0, alpha=0.001 / 9.0, beta=0.75, name='norm2') pool2 = tf.nn.max_pool(norm2, ksize=[1, 3, 3, 1], strides=[1, 2, 2, 1], padding='SAME', name='pool2') with tf.variable_scope('conv3') as scope: kernel = variable_with_weight_decay('weights', shape=[3, 3, 64, 128], stddev=5e-2, wd=0.0) conv = tf.nn.conv2d(pool2, kernel, [1, 1, 1, 1], padding='SAME') biases = variable_on_cpu('biases', [128], tf.constant_initializer(0.0)) pre_activation = tf.nn.bias_add(conv, biases) conv3 = tf.nn.relu(pre_activation, name=scope.name) #tf.summary.histogram('Convolution_layers/conv3', conv3) tf.summary.scalar('Convolution_layers/conv3', tf.nn.zero_fraction(conv3)) with tf.variable_scope('conv4') as scope: kernel = variable_with_weight_decay('weights', shape=[3, 3, 128, 128], stddev=5e-2, wd=0.0) conv = tf.nn.conv2d(conv3, kernel, [1, 1, 1, 1], padding='SAME') biases = variable_on_cpu('biases', [128], tf.constant_initializer(0.0)) pre_activation = tf.nn.bias_add(conv, biases) conv4 = tf.nn.relu(pre_activation, name=scope.name) #tf.summary.histogram('Convolution_layers/conv4', conv4) tf.summary.scalar('Convolution_layers/conv4', tf.nn.zero_fraction(conv4)) with tf.variable_scope('conv5') as scope: kernel = variable_with_weight_decay('weights', shape=[3, 3, 128, 128], stddev=5e-2, wd=0.0) conv = tf.nn.conv2d(conv4, kernel, [1, 1, 1, 1], padding='SAME') biases = variable_on_cpu('biases', [128], tf.constant_initializer(0.0)) pre_activation = tf.nn.bias_add(conv, biases) conv5 = tf.nn.relu(pre_activation, name=scope.name) #tf.summary.histogram('Convolution_layers/conv5', conv5) tf.summary.scalar('Convolution_layers/conv5', tf.nn.zero_fraction(conv5)) norm3 = tf.nn.lrn(conv5, 4, bias=1.0, alpha=0.001 / 9.0, beta=0.75, name='norm3') pool3 = tf.nn.max_pool(norm3, ksize=[1, 3, 3, 1], strides=[1, 2, 2, 1], padding='SAME', name='pool3') with tf.variable_scope('fully_connected1') as scope: reshape = tf.reshape(pool3, [-1, _RESHAPE_SIZE]) dim = reshape.get_shape()[1].value weights = variable_with_weight_decay('weights', shape=[dim, 384], stddev=0.04, wd=0.004) biases = variable_on_cpu('biases', [384], tf.constant_initializer(0.1)) local3 = tf.nn.relu(tf.matmul(reshape, weights) + biases, name=scope.name) #tf.summary.histogram('Fully connected layers/fc1', local3) tf.summary.scalar('Fully connected layers/fc1', tf.nn.zero_fraction(local3)) with tf.variable_scope('fully_connected2') as scope: weights = variable_with_weight_decay('weights', shape=[384, 192], stddev=0.04, wd=0.004) biases = variable_on_cpu('biases', [192], tf.constant_initializer(0.1)) local4 = tf.nn.relu(tf.matmul(local3, weights) + biases, name=scope.name) #tf.summary.histogram('Fully connected layers/fc2', local4) tf.summary.scalar('Fully connected layers/fc2', tf.nn.zero_fraction(local4)) with tf.variable_scope('output1') as scope: weights = variable_with_weight_decay('weights', [192, _NUM_CLASSES], stddev=1 / 192.0, wd=0.0) biases = variable_on_cpu('biases', [_NUM_CLASSES], tf.constant_initializer(0.0)) softmax_linear = tf.add(tf.matmul(local4, weights), biases, name=scope.name) #tf.summary.histogram('Fully connected layers/output', softmax_linear) tf.summary.scalar('Fully connected layers/output', tf.nn.zero_fraction(softmax_linear)) global_step = tf.Variable(initial_value=0, name='global_step', trainable=False) y_pred_cls = tf.argmax(softmax_linear, axis=1) return x, y, softmax_linear, global_step, y_pred_cls`

 

You can find my meta file here : https://drive.google.com/file/d/19g3nR_CFbqOUYF4rCPk-udhSozx35RJJ/view?usp=sharing

 

Thank you !

0 Kudos
5 Replies
Highlighted
Community Manager
19 Views

@AnnaBar I'm happy to help you debug this error. Thank you for providing your meta file, but I am unable to reproduce your issue without the index and weights file also. If you can provide those, that would be of great help. Thanks.

0 Kudos
Highlighted
Community Manager
19 Views

0 Kudos
Highlighted
Community Manager
19 Views

I found what caused the error. It was the operation tf.nn.zero_fraction()

0 Kudos
Highlighted
Community Manager
19 Views

@Annabar Removing any instances of tf.nn.zero_fraction() fixed the error?

0 Kudos
Highlighted
Community Manager
19 Views

Yes !

0 Kudos