Fine Tune
Hello ichsan, how to fine tune the dataset, which column we use to fine tune the dataset
Hello, since it was instruction dataset. So, the Question is labeled with "input" column and the answer is labeled with "output" column. There are so many ways for fine tune it. Three of them are using artidoro/Qlora, ludwig-ai/ludwig, or OpenAccess-AI-Collective/axolotl.
I want to build a model like a chatbot with this dataset, I follow the tutorial "fine tune question answering task", but I'm having difficulty building the model because the examples from the existing dataset have several columns such as the context column, while this dataset only has question and answer columns, what do you think? can you help me
This dataset doesn't have context column since it was only Q-A dataset. Maybe you can create your own context column. For example:
"You are a clever AI. Please answer correctly and factual."
thanks, I'll try it