TensorFlowOnSpark brings TensorFlow programs to Apache Spark clusters.
TFNode.start_cluster_server
, which is not required for tf.keras
and tf.estimator
.TFNode.export_saved_model
, which can be replaced by TF native APIs now.TFNodeContext.num_workers
to count master
, chief
, and worker
nodes.InputMode.TENSORFLOW
support for ML Pipelines, since the input data is always a Spark DataFrame for this API.HasMasterNode
and HasGraceSecs
params.export_fn
argument for Spark ML TFEstimator
(use TF export APIs instead).signature_def_key
and tag_set
for Spark ML TFModel
.TFModel
for TF2.x APIs.master
branch will be moving to TF 2.0 compatibility.tensorflow
as a dependency, in order to support other variants like tensorflow-gpu
or tf-nightly
.evaluator
node type in cluster (thanks to @bbshetty)util.single_node_env()
, which can be used to initialize the environment (HDFS compatibility + GPU allocation) for running a single-node instance of TensorFlow on the Spark driver.timeout
argument to TFCluster.shutdown()
(default is 3 days). This is intended to shutdown the Spark application in the event that any of the TF nodes hang for any reason. Set to -1 to disable timeout.tf.SparseTensor
support.__version__
to module.sys.path
to tensorboard search path.tf.estimator
.mnist/keras/mnist_mlp_estimator.py
with example of distributed/parallel inferencing via estimator.predict
.feed_timeout
argument to TFCluster.train()
for InputMode.SPARK.grace_secs
argument to TFCluster.shutdown()
.get_ip_address
(contributed by @viplav).assert
statements.TFCluster.shutdown()
tf.data
, add instructions, and misc code cleanup (from @yileic) tf.keras
apis