Use of accelerate

[glatq]

Hi! I am trying to use accelerate to boost the performance of the model using riemannian_sgd optimiser. Although, I am not able to figure out how to set it up properly. I tried to follow the tutorials on accelerate's page but with no luck. Do you have any specific instructions or an example so I can utilise my GPU for training the model? Thank you!!

[Bihaqo]

Hi! May I ask you why do you want to use this codebase? Both the idea and the code is a bit old, I wouldn't recommend using it for practical purposes

[glatq]

I am trying to replicate the training of the car dataset from the paper using Riemannian sgd for the purpose of a paper. So It would be really beneficial to make it work with accelerate so I can do faster trainings. Is there any updated codebase for that?

[Bihaqo]

I see, makes sense! With tensorflow GPU acceleration should be available out of the box, but you might need to wrap the highest level function in tf.function e.g.
@tf.function
def f():
do everything tf related
f()

good luck!