Closed doloresgarcia closed 8 months ago
I use the gatr model with a pytorch lightning wrapper for the training and a custom dataset. Using this model is resulting in a host memory leak, while other models do not result in that problem. Have you observed this in your trainings?
I use the gatr model with a pytorch lightning wrapper for the training and a custom dataset. Using this model is resulting in a host memory leak, while other models do not result in that problem. Have you observed this in your trainings?