Hi, I was reading the original paper and noticed that it mentions using nonlinear transformations for data augmentation during self-supervised pretraining. However, I couldn't find the corresponding implementation of this in this repository. I am interested in the details of the nonlinear transformations applied for data augmentation, and if possible, could you provide the relevant code and parameter settings as a reference?
Additionally, has there been any validation regarding the impact of this augmentation on model performance? Any insights on this would be greatly appreciated!
Thanks.
Hi, I was reading the original paper and noticed that it mentions using nonlinear transformations for data augmentation during self-supervised pretraining. However, I couldn't find the corresponding implementation of this in this repository. I am interested in the details of the nonlinear transformations applied for data augmentation, and if possible, could you provide the relevant code and parameter settings as a reference?
Additionally, has there been any validation regarding the impact of this augmentation on model performance? Any insights on this would be greatly appreciated!
Thanks.