Potentially related: #25138.

Gaurav is working on a proper replacement, it's not ready yet, for now please use compat.v1.

We don't know how to use compat.v1 to do this. If there is a way, would it be possible to answer the Stackoverflow question so that it would be possible for us to try TF 2? The implication from @allenlavoie was that no workaround existed.

@girving - If you use the upgrade script on your original TensorFlow model, tf.Session(config=...) should change to tf.compat.v1.Session(config=...) and work as expected.

If I want to experiment with writing "true TF 2.0 code", will tf.compat.v1.enable_eager_execution(config=...) have the desired effect? In other words, should I expect that we'll end up with a drop-in replacement for the above command?

@samsamoa: I've landed a fix to #25138. I'm continuing to work on additional config options. Could you please share which specific options are of interest to you?

@jaingaurav I use config.device_count['GPU'] and config.gpu_options.visible_device_list currently.

@samsamoa: I've landed a fix to #25138. I'm continuing to work on additional config options. Could you please share which specific options are of interest to you?

allow_growth appears to be an essential hack to fix some cudnn issue on rtx20xx devices.

config.gpu_options.allow_growth = True
config.log_device_placement = True # to log device placement (on which device the operation ran)

A number of new API were added in tf.config namespace to support this use case. Please let me know if there is anything we missed regarding this specific issue.

Looks like you can do this in TF 2.0 now like this:

gpu_devices = tf.config.experimental.list_physical_devices('GPU')
for device in gpu_devices:
    tf.config.experimental.set_memory_growth(device, True)

Check https://www.tensorflow.org/guide/gpu#limiting_gpu_memory_growth

@opcecco you are a lifesaver.
Thank you.

Looks like you can do this in TF 2.0 now like this:

gpu_devices = tf.config.experimental.list_physical_devices('GPU')
for device in gpu_devices:
    tf.config.experimental.set_memory_growth(device, True)

I have placed this right before defining a keras Model() instance and compiling it. My taskmanager yet still shows that indepdently all video card memory is taken. What am I doing wrong?

how do we add this in C++?

The C++ API still uses sessions, so use the session config as you would in 1.x.