# How to wait for GPU memory to be released when using TensorFlow in a ray remote function
**URL:** https://discuss.ray.io/t/how-to-wait-for-gpu-memory-to-be-released-when-using-tensorflow-in-a-ray-remote-function/13480
**Category:** Ray Core
**Created:** [January 19, 2024, 3:51pm UTC](https://discuss.ray.io/t/how-to-wait-for-gpu-memory-to-be-released-when-using-tensorflow-in-a-ray-remote-function/13480 "2024-01-19T15:51:52Z")
**Posts on this page:** 2
**Page:** 1
### Author:  [@EdoCha](https://discuss.ray.io/u/EdoCha)
#### Post date: [January 19, 2024, 3:51pm UTC](https://discuss.ray.io/t/how-to-wait-for-gpu-memory-to-be-released-when-using-tensorflow-in-a-ray-remote-function/13480/1 "2024-01-19T15:51:52Z")
Hello,
I’m defining a ray remote function that is doing some tensorflow stuff on the GPU. I saw that I need to use max\_calls=1 in the function decorator such that the worker thread is killed and GPU memory released after the function is complete. I do see this working fine if I wait a bit (e.g. 1 second) after the function call is done before I try to do some more stuff with tensorflow reallocating GPU memory. But if I don’t wait, then probably the worker is not killed yet when I try to do some more stuff and the memory isn’t released yet.
Now, waiting for an abitrary amount of time is definitelly not something one wants to do, what I want to do is wait until I get some kind of confirmation that the worker is killed.
How can I do that ?
Some pseudo code:
```auto
@ray.remote(num_cpus=1, num_gpus=1, max_calls=1)
def use_tf():
# use tensorflow to do some stuff and then return some results...
tf.model.load()
return 1
# call the ray remote which uses tensorflow and get the result from it
obj_ref=use_tf.remote()
result=ray.get(obj_ref)
# wait a bit for gpu memory to be released
# this is what I want to replace by some code that would wait until the worker that executed the use_tf() function is shutdown and GPU memory has been released instead...
time.wait(2)
# do some tensorflow stuff using the GPU
tf.model.load()
```
---
### Author:  [@sangcho](https://discuss.ray.io/u/sangcho)
#### Post date: [January 25, 2024, 12:31pm UTC](https://discuss.ray.io/t/how-to-wait-for-gpu-memory-to-be-released-when-using-tensorflow-in-a-ray-remote-function/13480/2 "2024-01-25T12:31:35Z")
Maybe use ray.available\_resources and check if gpus are available?
If you do other tensorflow thing in other worker, you can also enforce this by using num\_gpus=1