Remove the restriction of video memory occupied by tensorflow GPU

The author found that RTX3070 8G The graphics card , stay tensorflow Can only call 5488MB, Probably 5G about
Because of existence memory_limit The limitation of .

As shown in the figure ：

In order to lift this restriction , Play the most important role of graphics card in deep learning , Look for methods on the Internet , Finally found a solution
As shown in the following code ：

physical_gpus = tf.config.list_physical_devices("GPU")
tf.config.experimental.set_virtual_device_configuration(
    physical_gpus[0],
    [tf.config.experimental.VirtualDeviceConfiguration(memory_limit=6000)]
    #  Increase the video memory limit to 6000MB
)
logical_gpus = tf.config.list_logical_devices("GPU")

effect ：

You can see , Run again tensorflow when ,gpu Training can be used 6000MB, This value can be adjusted according to your own needs .