Add RAI checker for installed backends #137
Conversation
|
Thanks for catching this and taking care of it. My main thought on this is: if it's only used in tests... should it just be in I suppose it might be nice to have this functionality internally. I can think of at least one use case which is to be able to discern wether a user has built any backends at all when initializing an If we do want it internally, it's not necessarily a configuration option as much as it is a helper function. I'm personally leaning towards a helper function. thoughts?? |
|
@Spartee Yeah, I think we can move it to |
tests/backends/test_tf.py
Outdated
| try: | ||
| from tensorflow import keras | ||
|
|
||
| from smartsim.tf import freeze_model | ||
| except (ImportError, SmartSimError): | ||
| should_run = False | ||
| tf_availavble = False |
smartsim/_core/config/config.py
Outdated
| @@ -27,6 +27,7 @@ | |||
| import os | |||
| from functools import lru_cache | |||
| from pathlib import Path | |||
| from typing import List | |||
This PR adds data loaders which can be used with TensorFlow and PyTorch to train (or fit) a model. The loaders get the data from the Orchestrator and can be used when training in a distributed fashion (e.g with Horovod). [ committed by @al-rigazzi ] [ reviewed by @Spartee and @MattToast ]
This PR mainly adds
CONFIG.get_installed_backends, a property returning a list of installed Redis AI backends.This is used e.g. in tests, to check whether the corresponding test should be run. This also helps distinguishing between tests we cannot run because there is no backend, and tests we cannot run because the library is not installed (e.g.
test_freeze_modelwhich does not need Redis AI).Notice that for backend tests, we also check whether the module is (still) available, to prevent cases in which a user might have built RedisAI in another env.