► Keras 3 API documentation / Utilities / Keras configuration utilities

Keras configuration utilities

`version` function

keras.version()

Guides and examples using version

[source]

`clear_session` function

keras.utils.clear_session(free_memory=True)

Resets all state generated by Keras.

Keras manages a global state, which it uses to implement the Functional model-building API and to uniquify autogenerated layer names.

If you are creating many models in a loop, this global state will consume an increasing amount of memory over time, and you may want to clear it. Calling clear_session() releases the global state: this helps avoid clutter from old models and layers, especially when memory is limited.

Arguments

free_memory: Whether to call Python garbage collection. It's usually a good practice to call it to make sure memory used by deleted objects is immediately freed. However, it may take a few seconds to execute, so when using clear_session() in a short loop, you may want to skip it.

Example 1: calling clear_session() when creating models in a loop

for _ in range(100):
  # Without `clear_session()`, each iteration of this loop will
  # slightly increase the size of the global state managed by Keras
  model = keras.Sequential([
      keras.layers.Dense(10) for _ in range(10)])

for _ in range(100):
  # With `clear_session()` called at the beginning,
  # Keras starts with a blank state at each iteration
  # and memory consumption is constant over time.
  keras.backend.clear_session()
  model = keras.Sequential([
      keras.layers.Dense(10) for _ in range(10)])

Example 2: resetting the layer name generation counter

>>> layers = [keras.layers.Dense(10) for _ in range(10)]
>>> new_layer = keras.layers.Dense(10)
>>> print(new_layer.name)
dense_10
>>> keras.backend.clear_session()
>>> new_layer = keras.layers.Dense(10)
>>> print(new_layer.name)
dense

[source]

`enable_traceback_filtering` function

keras.config.enable_traceback_filtering()

Turn on traceback filtering.

Raw Keras tracebacks (also known as stack traces) involve many internal frames, which can be challenging to read through, while not being actionable for end users. By default, Keras filters internal frames in most exceptions that it raises, to keep traceback short, readable, and focused on what's actionable for you (your own code).

See also keras.config.disable_traceback_filtering() and keras.config.is_traceback_filtering_enabled().

If you have previously disabled traceback filtering via keras.config.disable_traceback_filtering(), you can re-enable it via keras.config.enable_traceback_filtering().

[source]

`disable_traceback_filtering` function

keras.config.disable_traceback_filtering()

Turn off traceback filtering.

See also keras.config.enable_traceback_filtering() and keras.config.is_traceback_filtering_enabled().

If you have previously disabled traceback filtering via keras.config.disable_traceback_filtering(), you can re-enable it via keras.config.enable_traceback_filtering().

Guides and examples using disable_traceback_filtering

Highly accurate boundaries segmentation using BASNet

[source]

`is_traceback_filtering_enabled` function

keras.config.is_traceback_filtering_enabled()

Check if traceback filtering is enabled.

See also keras.config.enable_traceback_filtering() and keras.config.disable_traceback_filtering().

If you have previously disabled traceback filtering via keras.config.disable_traceback_filtering(), you can re-enable it via keras.config.enable_traceback_filtering().

Returns

Boolean, True if traceback filtering is enabled, and False otherwise.

[source]

`enable_interactive_logging` function

keras.config.enable_interactive_logging()

Turn on interactive logging.

When interactive logging is enabled, Keras displays logs via stdout. This provides the best experience when using Keras in an interactive environment such as a shell or a notebook.

[source]

`disable_interactive_logging` function

keras.config.disable_interactive_logging()

Turn off interactive logging.

When interactive logging is disabled, Keras sends logs to absl.logging. This is the best option when using Keras in a non-interactive way, such as running a training or inference job on a server.

[source]

`is_interactive_logging_enabled` function

keras.config.is_interactive_logging_enabled()

Check if interactive logging is enabled.

To switch between writing logs to stdout and absl.logging, you may use keras.config.enable_interactive_logging() and keras.config.disable_interactive_logging().

Returns

Boolean, True if interactive logging is enabled, and False otherwise.

[source]

`enable_unsafe_deserialization` function

keras.config.enable_unsafe_deserialization()

Disables safe mode globally, allowing deserialization of lambdas.

[source]

`floatx` function

keras.config.floatx()

Return the default float type, as a string.

E.g. 'bfloat16', 'float16', 'float32', 'float64'.

Returns

String, the current default float type.

Example

>>> keras.config.floatx()
'float32'

[source]

`set_floatx` function

keras.config.set_floatx(value)

Set the default float dtype.

Note: It is not recommended to set this to "float16" for training, as this will likely cause numeric stability issues. Instead, mixed precision, which leverages a mix of float16 and float32. It can be configured by calling keras.mixed_precision.set_dtype_policy('mixed_float16').

Arguments

value: String; 'bfloat16', 'float16', 'float32', or 'float64'.

Examples

>>> keras.config.floatx()
'float32'

>>> keras.config.set_floatx('float64')
>>> keras.config.floatx()
'float64'

>>> # Set it back to float32
>>> keras.config.set_floatx('float32')

Raises

ValueError: In case of invalid value.

[source]

`image_data_format` function

keras.config.image_data_format()

Return the default image data format convention.

Returns

A string, either 'channels_first' or 'channels_last'.

Example

>>> keras.config.image_data_format()
'channels_last'

[source]

`set_image_data_format` function

keras.config.set_image_data_format(data_format)

Set the value of the image data format convention.

Arguments

data_format: string. 'channels_first' or 'channels_last'.

Examples

>>> keras.config.image_data_format()
'channels_last'

>>> keras.config.set_image_data_format('channels_first')
>>> keras.config.image_data_format()
'channels_first'

>>> # Set it back to `'channels_last'`
>>> keras.config.set_image_data_format('channels_last')

[source]

`epsilon` function

keras.config.epsilon()

Return the value of the fuzz factor used in numeric expressions.

Returns

A float.

Example

>>> keras.config.epsilon()
1e-07

[source]

`set_epsilon` function

keras.config.set_epsilon(value)

Set the value of the fuzz factor used in numeric expressions.

Arguments

value: float. New value of epsilon.

Examples

>>> keras.config.epsilon()
1e-07

>>> keras.config.set_epsilon(1e-5)
>>> keras.config.epsilon()
1e-05

>>> # Set it back to the default value.
>>> keras.config.set_epsilon(1e-7)

[source]

`backend` function

keras.config.backend()

Publicly accessible method for determining the current backend.

Returns

String, the name of the backend Keras is currently using. One of "tensorflow", "torch", or "jax".

Example

>>> keras.config.backend()
'tensorflow'

Guides and examples using backend

[source]

`set_backend` function

keras.config.set_backend(backend)

Reload the backend (and the Keras package).

Example

>>> import os
>>> os.environ["KERAS_BACKEND"] = "tensorflow"
>>>
>>> import keras
>>> from keras import ops
>>> type(ops.ones(()))
<class 'tensorflow.python.framework.ops.EagerTensor'>
>>>
>>> keras.config.set_backend("jax")
UserWarning: Using [`keras.config.set_backend`](/api/utils/config_utils#setbackend-function) is dangerous...
>>> del keras, ops
>>>
>>> import keras
>>> from keras import ops
>>> type(ops.ones(()))
<class 'jaxlib.xla_extension.ArrayImpl'>

⚠️ WARNING ⚠️: Using this function is dangerous and should be done carefully. Changing the backend will NOT convert the type of any already-instantiated objects. Thus, any layers / tensors / etc. already created will no longer be usable without errors. It is strongly recommended not to keep around any Keras-originated objects instances created before calling set_backend().

This includes any function or class instance that uses any Keras functionality. All such code needs to be re-executed after calling set_backend() and re-importing all imported keras modules.

[source]

`enable_flash_attention` function

keras.config.enable_flash_attention()

Enable flash attention.

Flash attention offers performance optimization for attention layers, making it especially useful for large language models (LLMs) that benefit from faster and more memory-efficient attention computations.

Once enabled, supported layers like MultiHeadAttention will attempt to use flash attention for faster computations. By default, this feature is enabled.

Note that enabling flash attention does not guarantee it will always be used. Typically, the inputs must be in float16 or bfloat16 dtype, and input layout requirements may vary depending on the backend.

[source]

`disable_flash_attention` function

keras.config.disable_flash_attention()

Disable flash attention.

Once disabled, supported layers like MultiHeadAttention will not use flash attention for faster computations.

[source]

`is_flash_attention_enabled` function

keras.config.is_flash_attention_enabled()

Checks whether flash attention is globally enabled in Keras.

Flash attention is a performance-optimized method for computing attention in large models, such as transformers, allowing for faster and more memory-efficient operations. This function checks the global Keras configuration to determine if flash attention is enabled for compatible layers (e.g., MultiHeadAttention).

Returns

False if disabled; otherwise, it indicates that it is enabled.

[source]

`is_nnx_enabled` function

keras.config.is_nnx_enabled()

Checks whether NNX specific features are enabled for the JAX backend.

Returns

bool: True if NNX backend features are enabled, False otherwise. Defaults to False.

[source]

`max_epochs` function

keras.config.max_epochs()

Get the maximum number of epochs for any call to fit.

Retrieves the limit on the number of epochs set by keras.config.set_max_epochs or the KERAS_MAX_EPOCHS environment variable.

Returns

The integer limit on the number of epochs or None, if no limit has been set.

[source]

`set_max_epochs` function

keras.config.set_max_epochs(max_epochs)

Limit the maximum number of epochs for any call to fit.

This will cap the number of epochs for any training run using model.fit(). This is purely for debugging, and can also be set via the KERAS_MAX_EPOCHS environment variable to quickly run a script without modifying its source.

Arguments

max_epochs: The integer limit on the number of epochs or None. If None, no limit is applied.

[source]

`max_steps_per_epoch` function

keras.config.max_steps_per_epoch()

Get the maximum number of steps for any call to fit/evaluate/predict.

Retrieves the limit on the number of epochs set by keras.config.set_max_steps_per_epoch or the KERAS_MAX_STEPS_PER_EPOCH environment variable.

Arguments

max_epochs: The integer limit on the number of epochs or None. If None, no limit is applied.

[source]

`set_max_steps_per_epoch` function

keras.config.set_max_steps_per_epoch(max_steps_per_epoch)

Limit the maximum number of steps for any call to fit/evaluate/predict.

This will cap the number of steps for single epoch of a call to fit(), evaluate(), or predict(). This is purely for debugging, and can also be set via the KERAS_MAX_STEPS_PER_EPOCH environment variable to quickly run a scrip without modifying its source.

Arguments