Get if memory growth is enabled for a PhysicalDevice.
tf.config.experimental.get_memory_growth(
device
)
If memory growth is enabled for a PhysicalDevice, the runtime initialization
will not allocate all memory on the device.
For example:
physical_devices = tf.config.list_physical_devices('GPU')
try:
tf.config.experimental.set_memory_growth(physical_devices[0], True)
assert tf.config.experimental.get_memory_growth(physical_devices[0])
except:
# Invalid device or cannot modify virtual devices once initialized.
pass
Args |
|---|
device
|
PhysicalDevice to query
|
Returns |
|---|
A boolean indicating the memory growth setting for the PhysicalDevice.
|
Raises |
|---|
ValueError
|
Invalid PhysicalDevice specified.
|
TensorFlow 1 version
View source on GitHub