DTypePolicy
class
keras.mixed_precision.DTypePolicy(name)
A dtype policy for a Keras layer.
A dtype policy determines a layer’s computation and variable dtypes. Each layer has a policy. Policies can be passed to the dtype
argument of layer constructors, or a global policy can be set with keras.config.set_dtype_policy
.
Arguments
- name: The policy name, which determines the compute and variable dtypes. Can be any dtype name, such as
"float32"
or"float64"
, which causes both the compute and variable dtypes will be that dtype. Can also be the string"mixed_float16"
or"mixed_bfloat16"
, which causes the compute dtype to befloat16
orbfloat16
and the variable dtype to befloat32
.
Typically you only need to interact with dtype policies when using mixed precision, which is the use of float16 or bfloat16 for computations and float32 for variables. This is why the term mixed_precision
appears in the API name. Mixed precision can be enabled by passing "mixed_float16"
or "mixed_bfloat16"
to keras.mixed_precision.set_dtype_policy()
.
>>> keras.config.set_dtype_policy("mixed_float16")
>>> layer1 = keras.layers.Dense(10)
>>> layer1.dtype_policy # layer1 will automatically use mixed precision
<DTypePolicy "mixed_float16">
>>> # Can optionally override layer to use float32
>>> # instead of mixed precision.
>>> layer2 = keras.layers.Dense(10, dtype="float32")
>>> layer2.dtype_policy
<DTypePolicy "float32">
>>> # Set policy back to initial float32.
>>> keras.config.set_dtype_policy('float32')
In the example above, passing dtype="float32"
to the layer is equivalent to passing dtype=keras.config.DTypePolicy("float32")
. In general, passing a dtype policy name to a layer is equivalent to passing the corresponding policy, so it is never necessary to explicitly construct a DTypePolicy
object.
dtype_policy
function
keras.mixed_precision.dtype_policy()
Returns the current default dtype policy object.
set_dtype_policy
function
keras.mixed_precision.set_dtype_policy(policy)
Sets the default dtype policy globally.
Example
>>> keras.config.set_dtype_policy("mixed_float16")
Leave a Reply