Splitnn requires NVFlare 2.4 #2003
-
|
In the requirements.txt of splitnn in advanced/vertical_federated_learning/cifar10_splitnn u need NVFlare 2.4 or higher. Should i use the 2.4.0rc1 or is there just a mistake in the requirements? |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment
-
|
@eshatkeinensinn thanks for the question! If you want to use the stable release version of NVFlare (2.3) you should use examples from 2.3 branch: https://github.com/NVIDIA/NVFlare/tree/2.3/examples/advanced/vertical_federated_learning |
Beta Was this translation helpful? Give feedback.
@eshatkeinensinn thanks for the question!
If you want to use the latest version of NVFlare (it is unstable though), you can ignore this requirements and instead do
pip install -e .[dev]in the cloned NVFlare repo.If you want to use the stable release version of NVFlare (2.3) you should use examples from 2.3 branch: https://github.com/NVIDIA/NVFlare/tree/2.3/examples/advanced/vertical_federated_learning