![]() |
![]() |
![]() |
![]() |
Overview
This guide goes beneath the surface of TensorFlow and Keras to demonstrate how TensorFlow works. If you instead want to immediately get started with Keras, check out the collection of Keras guides.
In this guide, you'll learn how TensorFlow allows you to make simple changes to your code to get graphs, how graphs are stored and represented, and how you can use them to accelerate your models.
This is a big-picture overview that covers how tf.function
allows you to switch from eager execution to graph execution. For a more complete specification of tf.function
, go to the Better performance with tf.function
guide.
What are graphs?
In the previous three guides, you ran TensorFlow eagerly. This means TensorFlow operations are executed by Python, operation by operation, and returning results back to Python.
While eager execution has several unique advantages, graph execution enables portability outside Python and tends to offer better performance. Graph execution means that tensor computations are executed as a TensorFlow graph, sometimes referred to as a tf.Graph
or simply a "graph."
Graphs are data structures that contain a set of tf.Operation
objects, which represent units of computation; and tf.Tensor
objects, which represent the units of data that flow between operations. They are defined in a tf.Graph
context. Since these graphs are data structures, they can be saved, run, and restored all without the original Python code.
This is what a TensorFlow graph representing a two-layer neural network looks like when visualized in TensorBoard:
The benefits of graphs
With a graph, you have a great deal of flexibility. You can use your TensorFlow graph in environments that don't have a Python interpreter, like mobile applications, embedded devices, and backend servers. TensorFlow uses graphs as the format for saved models when it exports them from Python.
Graphs are also easily optimized, allowing the compiler to do transformations like:
- Statically infer the value of tensors by folding constant nodes in your computation ("constant folding").
- Separate sub-parts of a computation that are independent and split them between threads or devices.
- Simplify arithmetic operations by eliminating common subexpressions.
There is an entire optimization system, Grappler, to perform this and other speedups.
In short, graphs are extremely useful and let your TensorFlow run fast, run in parallel, and run efficiently on multiple devices.
However, you still want to define your machine learning models (or other computations) in Python for convenience, and then automatically construct graphs when you need them.
Setup
Import some necessary libraries:
import tensorflow as tf
import timeit
from datetime import datetime
2022-12-14 04:02:43.649067: W tensorflow/compiler/xla/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libnvinfer.so.7'; dlerror: libnvinfer.so.7: cannot open shared object file: No such file or directory 2022-12-14 04:02:43.649163: W tensorflow/compiler/xla/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libnvinfer_plugin.so.7'; dlerror: libnvinfer_plugin.so.7: cannot open shared object file: No such file or directory 2022-12-14 04:02:43.649173: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Cannot dlopen some TensorRT libraries. If you would like to use Nvidia GPU with TensorRT, please make sure the missing libraries mentioned above are installed properly.
Taking advantage of graphs
You create and run a graph in TensorFlow by using tf.function
, either as a direct call or as a decorator. tf.function
takes a regular function as input and returns a Function
. A Function
is a Python callable that builds TensorFlow graphs from the Python function. You use a Function
in the same way as its Python equivalent.
# Define a Python function.
def a_regular_function(x, y, b):
x = tf.matmul(x, y)
x = x + b
return x
# `a_function_that_uses_a_graph` is a TensorFlow `Function`.
a_function_that_uses_a_graph = tf.function(a_regular_function)
# Make some tensors.
x1 = tf.constant([[1.0, 2.0]])
y1 = tf.constant([[2.0], [3.0]])
b1 = tf.constant(4.0)
orig_value = a_regular_function(x1, y1, b1).numpy()
# Call a `Function` like a Python function.
tf_function_value = a_function_that_uses_a_graph(x1, y1, b1).numpy()
assert(orig_value == tf_function_value)
On the outside, a Function
looks like a regular function you write using TensorFlow operations. Underneath, however, it is very different. A Function
encapsulates several tf.Graph
s behind one API (learn more in the Polymorphism section). That is how a Function
is able to give you the benefits of graph execution, like speed and deployability (refer to The benefits of graphs above).
tf.function
applies to a function and all other functions it calls:
def inner_function(x, y, b):
x = tf.matmul(x, y)
x = x + b
return x
# Use the decorator to make `outer_function` a `Function`.
@tf.function
def outer_function(x):
y = tf.constant([[2.0], [3.0]])
b = tf.constant(4.0)
return inner_function(x, y, b)
# Note that the callable will create a graph that
# includes `inner_function` as well as `outer_function`.
outer_function(tf.constant([[1.0, 2.0]])).numpy()
array([[12.]], dtype=float32)
If you have used TensorFlow 1.x, you will notice that at no time did you need to define a Placeholder
or tf.Session
.
Converting Python functions to graphs
Any function you write with TensorFlow will contain a mixture of built-in TF operations and Python logic, such as if-then
clauses, loops, break
, return
, continue
, and more. While TensorFlow operations are easily captured by a tf.Graph
, Python-specific logic needs to undergo an extra step in order to become part of the graph. tf.function
uses a library called AutoGraph (tf.autograph
) to convert Python code into graph-generating code.
def simple_relu(x):
if tf.greater(x, 0):
return x
else:
return 0
# `tf_simple_relu` is a TensorFlow `Function` that wraps `simple_relu`.
tf_simple_relu = tf.function(simple_relu)
print("First branch, with graph:", tf_simple_relu(tf.constant(1)).numpy())
print("Second branch, with graph:", tf_simple_relu(tf.constant(-1)).numpy())
First branch, with graph: 1 Second branch, with graph: 0
Though it is unlikely that you will need to view graphs directly, you can inspect the outputs to check the exact results. These are not easy to read, so no need to look too carefully!
# This is the graph-generating output of AutoGraph.
print(tf.autograph.to_code(simple_relu))
def tf__simple_relu(x): with ag__.FunctionScope('simple_relu', 'fscope', ag__.ConversionOptions(recursive=True, user_requested=True, optional_features=(), internal_convert_user_code=True)) as fscope: do_return = False retval_ = ag__.UndefinedReturnValue() def get_state(): return (do_return, retval_) def set_state(vars_): nonlocal do_return, retval_ (do_return, retval_) = vars_ def if_body(): nonlocal do_return, retval_ try: do_return = True retval_ = ag__.ld(x) except: do_return = False raise def else_body(): nonlocal do_return, retval_ try: do_return = True retval_ = 0 except: do_return = False raise ag__.if_stmt(ag__.converted_call(ag__.ld(tf).greater, (ag__.ld(x), 0), None, fscope), if_body, else_body, get_state, set_state, ('do_return', 'retval_'), 2) return fscope.ret(retval_, do_return)
# This is the graph itself.
print(tf_simple_relu.get_concrete_function(tf.constant(1)).graph.as_graph_def())
node { name: "x" op: "Placeholder" attr { key: "_user_specified_name" value { s: "x" } } attr { key: "dtype" value { type: DT_INT32 } } attr { key: "shape" value { shape { } } } } node { name: "Greater/y" op: "Const" attr { key: "dtype" value { type: DT_INT32 } } attr { key: "value" value { tensor { dtype: DT_INT32 tensor_shape { } int_val: 0 } } } } node { name: "Greater" op: "Greater" input: "x" input: "Greater/y" attr { key: "T" value { type: DT_INT32 } } } node { name: "cond" op: "StatelessIf" input: "Greater" input: "x" attr { key: "Tcond" value { type: DT_BOOL } } attr { key: "Tin" value { list { type: DT_INT32 } } } attr { key: "Tout" value { list { type: DT_BOOL type: DT_INT32 } } } attr { key: "_lower_using_switch_merge" value { b: true } } attr { key: "_read_only_resource_inputs" value { list { } } } attr { key: "else_branch" value { func { name: "cond_false_34" } } } attr { key: "output_shapes" value { list { shape { } shape { } } } } attr { key: "then_branch" value { func { name: "cond_true_33" } } } } node { name: "cond/Identity" op: "Identity" input: "cond" attr { key: "T" value { type: DT_BOOL } } } node { name: "cond/Identity_1" op: "Identity" input: "cond:1" attr { key: "T" value { type: DT_INT32 } } } node { name: "Identity" op: "Identity" input: "cond/Identity_1" attr { key: "T" value { type: DT_INT32 } } } library { function { signature { name: "cond_false_34" input_arg { name: "cond_placeholder" type: DT_INT32 } output_arg { name: "cond_identity" type: DT_BOOL } output_arg { name: "cond_identity_1" type: DT_INT32 } } node_def { name: "cond/Const" op: "Const" attr { key: "dtype" value { type: DT_BOOL } } attr { key: "value" value { tensor { dtype: DT_BOOL tensor_shape { } bool_val: true } } } } node_def { name: "cond/Const_1" op: "Const" attr { key: "dtype" value { type: DT_BOOL } } attr { key: "value" value { tensor { dtype: DT_BOOL tensor_shape { } bool_val: true } } } } node_def { name: "cond/Const_2" op: "Const" attr { key: "dtype" value { type: DT_INT32 } } attr { key: "value" value { tensor { dtype: DT_INT32 tensor_shape { } int_val: 0 } } } } node_def { name: "cond/Const_3" op: "Const" attr { key: "dtype" value { type: DT_BOOL } } attr { key: "value" value { tensor { dtype: DT_BOOL tensor_shape { } bool_val: true } } } } node_def { name: "cond/Identity" op: "Identity" input: "cond/Const_3:output:0" attr { key: "T" value { type: DT_BOOL } } } node_def { name: "cond/Const_4" op: "Const" attr { key: "dtype" value { type: DT_INT32 } } attr { key: "value" value { tensor { dtype: DT_INT32 tensor_shape { } int_val: 0 } } } } node_def { name: "cond/Identity_1" op: "Identity" input: "cond/Const_4:output:0" attr { key: "T" value { type: DT_INT32 } } } ret { key: "cond_identity" value: "cond/Identity:output:0" } ret { key: "cond_identity_1" value: "cond/Identity_1:output:0" } attr { key: "_construction_context" value { s: "kEagerRuntime" } } arg_attr { key: 0 value { attr { key: "_output_shapes" value { list { shape { } } } } } } } function { signature { name: "cond_true_33" input_arg { name: "cond_identity_1_x" type: DT_INT32 } output_arg { name: "cond_identity" type: DT_BOOL } output_arg { name: "cond_identity_1" type: DT_INT32 } } node_def { name: "cond/Const" op: "Const" attr { key: "dtype" value { type: DT_BOOL } } attr { key: "value" value { tensor { dtype: DT_BOOL tensor_shape { } bool_val: true } } } } node_def { name: "cond/Identity" op: "Identity" input: "cond/Const:output:0" attr { key: "T" value { type: DT_BOOL } } } node_def { name: "cond/Identity_1" op: "Identity" input: "cond_identity_1_x" attr { key: "T" value { type: DT_INT32 } } } ret { key: "cond_identity" value: "cond/Identity:output:0" } ret { key: "cond_identity_1" value: "cond/Identity_1:output:0" } attr { key: "_construction_context" value { s: "kEagerRuntime" } } arg_attr { key: 0 value { attr { key: "_output_shapes" value { list { shape { } } } } } } } } versions { producer: 1286 min_consumer: 12 }
Most of the time, tf.function
will work without special considerations. However, there are some caveats, and the tf.function
guide can help here, as well as the complete AutoGraph reference.
Polymorphism: one Function
, many graphs
A tf.Graph
is specialized to a specific type of inputs (for example, tensors with a specific dtype
or objects with the same id()
).
Each time you invoke a Function
with a set of arguments that can't be handled by any of its existing graphs (such as arguments with new dtypes
or incompatible shapes), Function
creates a new tf.Graph
specialized to those new arguments. The type specification of a tf.Graph
's inputs is known as its input signature or just a signature. For more information regarding when a new tf.Graph
is generated and how that can be controlled, go to the Rules of tracing section of the Better performance with tf.function
guide.
The Function
stores the tf.Graph
corresponding to that signature in a ConcreteFunction
. A ConcreteFunction
is a wrapper around a tf.Graph
.
@tf.function
def my_relu(x):
return tf.maximum(0., x)
# `my_relu` creates new graphs as it observes more signatures.
print(my_relu(tf.constant(5.5)))
print(my_relu([1, -1]))
print(my_relu(tf.constant([3., -3.])))
tf.Tensor(5.5, shape=(), dtype=float32) tf.Tensor([1. 0.], shape=(2,), dtype=float32) tf.Tensor([3. 0.], shape=(2,), dtype=float32)
If the Function
has already been called with that signature, Function
does not create a new tf.Graph
.
# These two calls do *not* create new graphs.
print(my_relu(tf.constant(-2.5))) # Signature matches `tf.constant(5.5)`.
print(my_relu(tf.constant([-1., 1.]))) # Signature matches `tf.constant([3., -3.])`.
tf.Tensor(0.0, shape=(), dtype=float32) tf.Tensor([0. 1.], shape=(2,), dtype=float32)
Because it's backed by multiple graphs, a Function
is polymorphic. That enables it to support more input types than a single tf.Graph
could represent, and to optimize each tf.Graph
for better performance.
# There are three `ConcreteFunction`s (one for each graph) in `my_relu`.
# The `ConcreteFunction` also knows the return type and shape!
print(my_relu.pretty_printed_concrete_signatures())
my_relu(x) Args: x: float32 Tensor, shape=() Returns: float32 Tensor, shape=() my_relu(x=[1, -1]) Returns: float32 Tensor, shape=(2,) my_relu(x) Args: x: float32 Tensor, shape=(2,) Returns: float32 Tensor, shape=(2,)
Using tf.function
So far, you've learned how to convert a Python function into a graph simply by using tf.function
as a decorator or wrapper. But in practice, getting tf.function
to work correctly can be tricky! In the following sections, you'll learn how you can make your code work as expected with tf.function
.
Graph execution vs. eager execution
The code in a Function
can be executed both eagerly and as a graph. By default, Function
executes its code as a graph:
@tf.function
def get_MSE(y_true, y_pred):
sq_diff = tf.pow(y_true - y_pred, 2)
return tf.reduce_mean(sq_diff)
y_true = tf.random.uniform([5], maxval=10, dtype=tf.int32)
y_pred = tf.random.uniform([5], maxval=10, dtype=tf.int32)
print(y_true)
print(y_pred)
tf.Tensor([6 6 1 7 5], shape=(5,), dtype=int32) tf.Tensor([3 1 6 9 6], shape=(5,), dtype=int32)
get_MSE(y_true, y_pred)
<tf.Tensor: shape=(), dtype=int32, numpy=12>
To verify that your Function
's graph is doing the same computation as its equivalent Python function, you can make it execute eagerly with tf.config.run_functions_eagerly(True)
. This is a switch that turns off Function
's ability to create and run graphs, instead of executing the code normally.
tf.config.run_functions_eagerly(True)
get_MSE(y_true, y_pred)
<tf.Tensor: shape=(), dtype=int32, numpy=12>
# Don't forget to set it back when you are done.
tf.config.run_functions_eagerly(False)
However, Function
can behave differently under graph and eager execution. The Python print
function is one example of how these two modes differ. Let's check out what happens when you insert a print
statement to your function and call it repeatedly.
@tf.function
def get_MSE(y_true, y_pred):
print("Calculating MSE!")
sq_diff = tf.pow(y_true - y_pred, 2)
return tf.reduce_mean(sq_diff)
Observe what is printed:
error = get_MSE(y_true, y_pred)
error = get_MSE(y_true, y_pred)
error = get_MSE(y_true, y_pred)
Calculating MSE!
Is the output surprising? get_MSE
only printed once even though it was called three times.
To explain, the print
statement is executed when Function
runs the original code in order to create the graph in a process known as "tracing" (refer to the Tracing section of the tf.function
guide. Tracing captures the TensorFlow operations into a graph, and print
is not captured in the graph. That graph is then executed for all three calls without ever running the Python code again.
As a sanity check, let's turn off graph execution to compare:
# Now, globally set everything to run eagerly to force eager execution.
tf.config.run_functions_eagerly(True)
# Observe what is printed below.
error = get_MSE(y_true, y_pred)
error = get_MSE(y_true, y_pred)
error = get_MSE(y_true, y_pred)
Calculating MSE! Calculating MSE! Calculating MSE!
tf.config.run_functions_eagerly(False)
print
is a Python side effect, and there are other differences that you should be aware of when converting a function into a Function
. Learn more in the Limitations section of the Better performance with tf.function
guide.
Non-strict execution
Graph execution only executes the operations necessary to produce the observable effects, which includes:
- The return value of the function
- Documented well-known side-effects such as:
- Input/output operations, like
tf.print
- Debugging operations, such as the assert functions in
tf.debugging
- Mutations of
tf.Variable
- Input/output operations, like
This behavior is usually known as "Non-strict execution", and differs from eager execution, which steps through all of the program operations, needed or not.
In particular, runtime error checking does not count as an observable effect. If an operation is skipped because it is unnecessary, it cannot raise any runtime errors.
In the following example, the "unnecessary" operation tf.gather
is skipped during graph execution, so the runtime error InvalidArgumentError
is not raised as it would be in eager execution. Do not rely on an error being raised while executing a graph.
def unused_return_eager(x):
# Get index 1 will fail when `len(x) == 1`
tf.gather(x, [1]) # unused
return x
try:
print(unused_return_eager(tf.constant([0.0])))
except tf.errors.InvalidArgumentError as e:
# All operations are run during eager execution so an error is raised.
print(f'{type(e).__name__}: {e}')
tf.Tensor([0.], shape=(1,), dtype=float32)
@tf.function
def unused_return_graph(x):
tf.gather(x, [1]) # unused
return x
# Only needed operations are run during graph execution. The error is not raised.
print(unused_return_graph(tf.constant([0.0])))
tf.Tensor([0.], shape=(1,), dtype=float32)
tf.function
best practices
It may take some time to get used to the behavior of Function
. To get started quickly, first-time users should play around with decorating toy functions with @tf.function
to get experience with going from eager to graph execution.
Designing for tf.function
may be your best bet for writing graph-compatible TensorFlow programs. Here are some tips:
- Toggle between eager and graph execution early and often with
tf.config.run_functions_eagerly
to pinpoint if/ when the two modes diverge. - Create
tf.Variable
s outside the Python function and modify them on the inside. The same goes for objects that usetf.Variable
, liketf.keras.layers
,tf.keras.Model
s andtf.keras.optimizers
. - Avoid writing functions that depend on outer Python variables, excluding
tf.Variable
s and Keras objects. Learn more in Depending on Python global and free variables of thetf.function
guide. - Prefer to write functions which take tensors and other TensorFlow types as input. You can pass in other object types but be careful! Learn more in Depending on Python objects of the
tf.function
guide. - Include as much computation as possible under a
tf.function
to maximize the performance gain. For example, decorate a whole training step or the entire training loop.
Seeing the speed-up
tf.function
usually improves the performance of your code, but the amount of speed-up depends on the kind of computation you run. Small computations can be dominated by the overhead of calling a graph. You can measure the difference in performance like so:
x = tf.random.uniform(shape=[10, 10], minval=-1, maxval=2, dtype=tf.dtypes.int32)
def power(x, y):
result = tf.eye(10, dtype=tf.dtypes.int32)
for _ in range(y):
result = tf.matmul(x, result)
return result
print("Eager execution:", timeit.timeit(lambda: power(x, 100), number=1000), "seconds")
Eager execution: 4.083200226999907 seconds
power_as_graph = tf.function(power)
print("Graph execution:", timeit.timeit(lambda: power_as_graph(x, 100), number=1000), "seconds")
Graph execution: 0.7515849150004215 seconds
tf.function
is commonly used to speed up training loops, and you can learn more about it in the _Speeding-up your training step with tf.function
_ section of the Writing a training loop from scratch with Keras guide.
Performance and trade-offs
Graphs can speed up your code, but the process of creating them has some overhead. For some functions, the creation of the graph takes more time than the execution of the graph. This investment is usually quickly paid back with the performance boost of subsequent executions, but it's important to be aware that the first few steps of any large model training can be slower due to tracing.
No matter how large your model, you want to avoid tracing frequently. The tf.function
guide discusses how to set input specifications and use tensor arguments to avoid retracing in the Controlling retracing section. If you find you are getting unusually poor performance, it's a good idea to check if you are retracing accidentally.
When is a Function
tracing?
To figure out when your Function
is tracing, add a print
statement to its code. As a rule of thumb, Function
will execute the print
statement every time it traces.
@tf.function
def a_function_with_python_side_effect(x):
print("Tracing!") # An eager-only side effect.
return x * x + tf.constant(2)
# This is traced the first time.
print(a_function_with_python_side_effect(tf.constant(2)))
# The second time through, you won't see the side effect.
print(a_function_with_python_side_effect(tf.constant(3)))
Tracing! tf.Tensor(6, shape=(), dtype=int32) tf.Tensor(11, shape=(), dtype=int32)
# This retraces each time the Python argument changes,
# as a Python argument could be an epoch count or other
# hyperparameter.
print(a_function_with_python_side_effect(2))
print(a_function_with_python_side_effect(3))
Tracing! tf.Tensor(6, shape=(), dtype=int32) Tracing! tf.Tensor(11, shape=(), dtype=int32)
New Python arguments always trigger the creation of a new graph, hence the extra tracing.
Next steps
You can learn more about tf.function
on the API reference page and by following the Better performance with tf.function
guide.