3 ms·
Fine-tuning and distillation used to mean two different things. Now not so much anymore, I see.
by moralestapia 3mo ago
Fine-tuning and distillation used to mean two different things. Now not so much anymore, I see.
- bizzletk 3mo agoCan you explain the difference?
- nerdsniper 3mo agoFine-tuning is done against a dataset. Distillation is done against a model.
- verdverm 3mo agoDistillation is used to build part of a data set for fine-tuning (loosely interpreted). Advanced model traces are useless if you don't have a base model that is good enough to be improved by them.
- porridgeraisin 3mo agoDistillation originally meant matching the distribution of the student model to the teacher model using something like a KL divergence. When you instead fine-tune the student on the samples from the teacher, which is what people mean by distillation today, you are in effect doing a monte-carlo version of the same thing. While in theory this is higher variance, given modern setups where the student and teacher are both large and are RLd heavily (leading to a sharp teacher distribution), and given that you typically use lots and lots of samples, it ends up OK.
- sapienskid 3mo agoHow to get access?