Version: Next
Accessing TensorBoard
Enabling TensorBoard
You can add TensorBoard as a sidecar to any Workflow template as follows:
templates:
- name: my-template
container:
...
sidecars:
- name: tensorboard
# Use Tensorflow docker image
image: tensorflow/tensorflow:2.3.0
command:
- sh
- '-c'
# Indicates that this is a interactive/visualization sidecar
env:
- name: ONEPANEL_INTERACTIVE_SIDECAR
value: 'true'
args:
# Set <path> to where your main container is writing TensorBoard logs
- tensorboard --logdir <path>
ports:
- containerPort: 6006
name: tensorboard
Complete example
note
This example is also available in the application in Workflows > Workflow Templates > TensorFlow Training.
arguments:
parameters:
- name: epochs
value: '10'
entrypoint: main
templates:
- name: main
dag:
tasks:
- name: train-model
template: tf-dense
- name: tf-dense
script:
image: tensorflow/tensorflow:2.3.0
command:
- python
- '-u'
source: |
import tensorflow as tf
import datetime
mnist = tf.keras.datasets.mnist
(x_train, y_train),(x_test, y_test) = mnist.load_data()
x_train, x_test = x_train / 255.0, x_test / 255.0
def create_model():
return tf.keras.models.Sequential([
tf.keras.layers.Flatten(input_shape=(28, 28)),
tf.keras.layers.Dense(512, activation='relu'),
tf.keras.layers.Dropout(0.2),
tf.keras.layers.Dense(10, activation='softmax')
])
model = create_model()
model.compile(optimizer='adam',
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])
# Write logs to /mnt/output
log_dir = "/mnt/output/logs/" + datetime.datetime.now().strftime("%Y%m%d-%H%M%S")
tensorboard_callback = tf.keras.callbacks.TensorBoard(log_dir=log_dir, histogram_freq=1)
history = model.fit(x=x_train,
y=y_train,
epochs={{workflow.parameters.epochs}},
validation_data=(x_test, y_test),
callbacks=[tensorboard_callback])
volumeMounts:
# TensorBoard sidecar will automatically mount this volume
- name: output
mountPath: /mnt/output
sidecars:
- name: tensorboard
image: tensorflow/tensorflow:2.3.0
command:
- sh
- '-c'
env:
- name: ONEPANEL_INTERACTIVE_SIDECAR
value: 'true'
args:
# Read logs from /mnt/output - this directory is auto-mounted from volumeMounts
- tensorboard --logdir /mnt/output/
ports:
- containerPort: 6006
name: tensorboard
volumeClaimTemplates:
# Provision a volume that can be shared between main container and TensorBoard side car
- metadata:
name: output
spec:
accessModes: [ "ReadWriteOnce" ]
resources:
requests:
storage: 2Gi
Launching TensorBoard
Once a Workflow task is running, you can access its TensorBoard sidecar by clicking on the task and then clicking Outputs and then Open TensorBoard in the task panel:
