def show_graph(graph_def, max_const_size=32):
"""Visualize TensorFlow graph."""
graph_file = open('graph.html', 'w')
if hasattr(graph_def, 'as_graph_def'):
graph_def = graph_def.as_graph_def()
strip_def = strip_consts(graph_def, max_const_size=max_const_size)
code = """
<script>
function load() {{
document.getElementById("{id}").pbtxt = {data};
}}
</script>
<link rel="import" href="https://tensorboard.appspot.com/tf-graph-basic.build.html"
onload=load()>
<div style="height:600px">
<tf-graph-basic id="{id}"><tf-graph-basic>
</div>
""".format(data=repr(str(strip_def)), id='graph' + str(np.random.rand()))
iframe = """
<iframe seamless style="width:1800px;height:1620px;border:0" srcdoc="{}"></iframe>
""".format(code.replace('"', '"'))
display(HTML(iframe))
with graph_file as g:
g.write(iframe)
filename = os.path.join(PROJECT_DIR, 'graph.html')
webbrowser.open_new_tab(filename)
# Use Graphviz to view graph
graph = tf.get_default_graph()
tfg.board(graph, depth=2).view()
import tensorflow as tf
# Graph
import tfgraphviz as tfg
# Declare Functions
def relu(X):
# Rectified Linear Unit
# h_{w,b} (X) = max ( X \cdot w + b, 0 )
with tf.name_scope("relu"):
w_shape = (int(X.get_shape()[1]), 1)
w = tf.Variable(tf.random_normal(w_shape), name="weights")
b = tf.Variable(0.0, name="bias")
z = tf.add(tf.matmul(X, w), b, name="z")
return tf.maximum(z, 0, name="relu")
# Create nodes
n_features = 3
X = tf.placeholder(tf.float32, shape=(None, n_features), name="X")
relus = [relu(X) for i in range(5)] # Compute multiple ReLU
output = tf.add_n(relus, name="output") # Sum a list of tensors
# Write out files
file_writer = tf.summary.FileWriter("logs/relu1", tf.get_default_graph())
file_writer.close()
# Generate Graph
tfg.board(tf.get_default_graph()).view()
#!/usr/bin/env python import tensorflow as tf import tfgraphviz as tfg a = tf.constant(1, name="a") b = tf.constant(2, name="b") c = tf.add(a, b, name="add") g = tfg.board(tf.get_default_graph()) g.view()
X_train = X_train.astype(np.float32).reshape(-1, 28 * 28) / 255.0
X_test = X_test.astype(np.float32).reshape(-1, 28 * 28) / 255.0
y_train = y_train.astype(np.int32)
y_test = y_test.astype(np.int32)
X_valid, X_train = X_train[:5000], X_train[5000:]
y_valid, y_train = y_train[:5000], y_train[5000:]
# Load Model
saver = tf.train.import_meta_graph("./tf_mnist_model_final.ckpt.meta")
# Display opeartions in model from graph
for op in tf.get_default_graph().get_operations():
print(op.name)
# Display graph
tfg.board(tf.get_default_graph(), depth=3).view()
# Getting tensors by name
X = tf.get_default_graph().get_tensor_by_name("X:0")
y = tf.get_default_graph().get_tensor_by_name("y:0")
accuracy = tf.get_collection().get_tensor_by_name("eval/acc:0")
training_op = tf.get_default_graph().get_operation_by_name("GradientDescent")
# To restore training model operations from imported graph
X, y, accuracy, training_op = tf.get_collection("my_important_ops")
# To restore model
with tf.Session() as sess:
saver.restore(sess, "./tf_mnist_model_final.ckpt")
# Continue training model
def graph2pdf(sess, directory, **kw):
print('Saving graph PDF in', directory, end=' ... '); sys.stdout.flush()
g = tfg.board(sess.graph, **kw)
g.render(filename='graph', directory=directory)
print('done.')