Create a weighted loss function to handle imbalance?

[John6666]

There don’t seem to be many examples…

SFTTrainer Loss function

I have a couple of questions: how to know the loss function used by default for SFTTrainer for a given model and how to alter it? For training an LLM the loss function is computed on the whole concatenated prompts, how to alter this and make loss function only compute on the output prompts

How to specify the loss function when finetuning a model using the Huggingface TFTrainer Class?