You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I have been studying LoRA recently and I noticed that during pre-training, the word vectors change as the training progresses. However, what about when using LoRA for fine-tuning? Do the word vectors still change, or is it only the attention weights.